mdja understanding the horizontal gaze nystagmus

Included in Materials:

Handout, The Horizontal Gaze Nystagmus (HGN) Test by Karl Citek, OD, PhD, FAAO ............2

Article, Nystagmus Testing in Intoxicated Individuals by Karl Citek, OD, PhD, Bret Ball, OD, and Dale A. Rutledge, Lieutenant (2003) ............................................................................8

Article, Sleep Deprivation Does Not Mimic Alcohol Intoxication on Field Sobriety Testing by Karl Citek, OD, PhD; Ashlee D. Elmont, OD; Christopher L. Jons, OD; Chad J. Krezelok, OD; Joseph D. Neron, OD; Timothy A. Plummer; and Timothy Tannenbaum (2011) ....24

Understanding the Horizontal Gaze Nystagmus

Dr. Karl Citek, MS, OD, PhD, FAAO Pacific University College of Optometry

August 11-12, 2020

1

The Horizontal Gaze Nystagmus (HGN) Test Karl Citek, OD, PhD, FAAO

Professor of Optometry Pacific University College of Optometry

Forest Grove, Oregon (503) 352-2126

[email protected]

I. Purpose of the Standardized Field Sobriety Tests (SFSTs)

A. SFSTs are screening tests that allow an officer to articulate probable cause to arrest a

driver whom s/he suspects of being impaired.

B. Most officers do not conduct SFSTs on drivers for whom they do not have reasonable

suspicion of impairment at the initial contact of a traffic stop. Therefore, sober drivers

who may have eye movement or balance problems typically will not be mistaken as

intoxicated since there would be no reason for an officer to conduct SFSTs, unless the

officer suspects impairment/intoxication for other reasons.

C. SFSTs correlate to physical impairment caused by intoxication. However, because there

literally are too many factors to consider, they do not, and never were intended to,

correlate to driving impairment. Yet, they quickly and efficiently assess physical and

mental skills similar to those necessary in order to safely operate a motor vehicle.

II. Sequelae of intoxication

A. In terms of physical signs and physiological indicators, intoxication affects the

functioning of the Central Nervous System (CNS), specifically, parts of the brainstem

and cerebellum.

B. Functions potentially affected by intoxication

1. Eye movements and/or pupil responses, depending on the intoxicant(s)

2. Speech

3. Gross and fine motor skills, e.g., movement of one limb and movement of fingers,

respectively

4. Coordinated motor skills, e.g., walking (deficit is termed gait ataxia)

C. Even to this day, we do not know the precise mechanism by which alcohol and many

drugs actually cause these changes; that doesn’t mean they don’t happen!

MDJA Annual Conference | August 11-12, 2020

2

III. HGN Health Questions

A. Glasses or contact lenses?

1. Suspect’s spectacles are removed to allow the officer an unobstructed view of the

suspect’s eyes. The suspect’s possible inability to see the stimulus clearly does not

invalidate the test.

2. Suspect’s contact lenses (CLs) are kept in place, since the suspect is not asked to

make any eye movements that he/she would not normally make. About 10% of CL

wearers use gas permeable (rigid) lenses, the remainder use soft lenses. All CLs

potentially reduce nystagmus, regardless of its origin or cause, probably via a

biofeedback mechanism.

B. Blind in one eye, or other eye problems?

1. Pilot studies show that monocular (one-eyed) individuals show the expected clues in

the remaining eye.

2. Individuals with one blind eye, or one eye with severely reduced vision (e.g.,

significant amblyopia), show the expected clues for all tests in which the good eye

can see the stimulus (usually everything except Distinct & Sustained Nystagmus at

Maximum Deviation [see below] for the non-seeing eye).

C. Any eye or head injuries or other medical problems?

1. Cerebral vascular accident (CVA or stroke), hypoglycemia (insulin shock) in

diabetes, heart attack, etc., should be evident as impairment but not intoxication to a

properly-trained officer.

2. Vestibular system or joint/muscle problems that could affect balance.

D. Other possible factors

1. Sleep deprivation does not cause or exacerbate any clues/indicators on any SFST in

the absence of intoxication (research study by Citek et al. 2011).

2. There is no evidence that anxiety, time of day, slight differences in weather

conditions, etc. can cause or exacerbate any clues/indicators on any SFST in the

absence of intoxication.


3

IV. Horizontal Gaze Nystagmus (HGN) Test

A. Pre-Test Checks

1. Check for equal pupil sizes

a. Up to 35% of normal, sober individuals have a difference in pupil sizes of 0.5

mm or more, but fewer than about 10% have a difference of 1.0 mm or more.

b. Most people with any difference in pupil sizes are aware of it.

c. Recent onset, or no prior knowledge, can indicate a recent head injury or stroke.

2. Check for equal tracking & “resting nystagmus”

a. Checks that eyes can move together and have full range of motion.

b. Unequal tracking indicates presence of strabismus (misalignment of the eyes) or

other neurological problem.

c. Nystagmus in primary gaze (“resting nystagmus”) most commonly is a medical

condition, but, when other obvious physical and physiological indicators of drug

use are apparent, could indicate intoxication by PCP.

B. Sub-tests are conducted in the order in which they are expected to appear with increasing

levels of intoxication.

1. Lack of Smooth Pursuit

a. About 10% of normal, sober individuals have lack of smooth pursuit, but they

will not be tested at a traffic stop unless the officer suspects impairment.

b. Research repeatedly has shown that normal, sober individuals should be able to

follow a stimulus moving across the visual field at a speed of about 30 degrees

per second, such that it will take the officer about 2 seconds (± 0.5 second) to

move a stimulus from center to either side, and about 4 seconds (± 1 second) to

move it from one side to the other.

c. If smooth pursuit eye movements are not present, for whatever reason, the

individual must make saccades (a.k.a. saccadic eye movements), which is the

expected finding for this test, or move the entire head.

d. Lack of Smooth Pursuit can be evident at BACs as low as .02-.03, making this

an effective indicator for Commercial Drivers (.04 Federal per se limit) and

Minors (.02 or zero tolerance, depending on jurisdiction) when other physical

and physiological evidence of intoxication is present.


4

e. Moving or positioning the stimulus inappropriately can affect the observation of

clues:

i. If the officer moves the stimulus too quickly, s/he could induce Lack of

Smooth Pursuit but will more likely miss observing the clue.

ii. If the officer holds the stimulus closer than 12 inches, s/he will be more

likely to observe Lack of Smooth Pursuit but will be at an unsafe close

distance to the suspect.

iii. If the officer holds the stimulus further than 15 inches, s/he may have

difficulty observing the suspect’s eyes and will no longer have proper

control of the suspect.

2. Distinct & Sustained Nystagmus at Maximum Deviation

a. Distinguished from endpoint nystagmus (EN), which is a phenomenon exhibited

by 50-60% of normal, sober individuals.

i. EN usually is of small amplitude; therefore, not distinct; and

ii. EN usually dissipates within 1-2 sec; therefore, not sustained.

b. Direction of nystagmus (given by the direction of the fast phase of the eye

movement) is always in the direction of gaze.

3. Onset of Nystagmus Prior to 45 Degrees

a. Referred to as gaze-evoked, gaze-induced, or simply gaze nystagmus if it is

neurological in origin. However, if it is of neurological origin, the earlier clues

generally will NOT be present, nor will other physical evidence of intoxication.

b. Stimulus is moved from center to the side at about half the speed of the check

for Lack of Smooth Pursuit, about 10-15 degrees per second, such that it would

take about 4 seconds to reach 45 degrees.

c. 45 degrees may be estimated from the position of the stimulus with respect to

the suspect’s shoulder.

i. For persons of average body build, distance from center to edge of shoulder

is about 10 (±1) inches.

ii. If the stimulus is held 12-15 inches from the eyes, the officer typically can

go several inches beyond the edge of the shoulder and still be within 45

degrees.


5

d. Research repeatedly has shown that the angle of onset of nystagmus (AON)

estimates BAC, if alcohol is the only intoxicant, by the approximate formula

referred to as Tharp’s equation,

BAC ≈ (50 – AON) / 100

i. This formula is most accurate for BACs between .10 and .20.

ii. AON = 45 degrees actually correlates with BAC of .08.

iii. BAC > .20 can show “immediate” AON and/or other very definite and

obvious signs of intoxication.

iv. Even though officers likely will not be able to testify to this assessment,

they are taught that a small AON without a concomitant high BAC

indicates the presence of drugs other than or in addition to alcohol.

4. Scoring

a. One point for the presence of each of three indicators in each eye, for a

maximum of 6 points.

b. At low to moderate BAC levels (< .08), about 10% of subjects show an odd

number of clues, likely due to a slight, clinically-insignificant left-right

asymmetry.

c. Results are not expected to vary or change with one’s experience with alcohol or

drugs. However, some individuals show all 6 clues at BACs below .08 (without

the presence of other intoxicants that also could affect the eyes), while others

show fewer than 4 (or even zero) clues at BACs above .10.

V. Vertical Gaze Nystagmus (VGN) Test

A. Originally part of the Drug Recognition Expect (DRE, a.k.a. Drug Evaluation and

Classification Program, DECP) protocol, the VGN Test was added to the SFST protocol

in 2002.

B. To be conducted after the HGN Test, but not part of the HGN Test or scoring.

C. Expect to observe distinct & sustained vertical nystagmus in maximum upgaze.

D. Unlike the HGN Test clues, presence of vertical nystagmus does vary with one’s

experience with alcohol or drugs.


6

E. Presence indicates a high level of intoxication for that individual, regardless of the actual

BAC or amount of drug(s) in the body.

F. If the officer does not conduct the VGN Test, this does not invalidate the findings on the

HGN Test.

VI. (Non-)Effects of Environmental Conditions

A. HGN and VGN tests, and others, can be properly performed with suspects standing,

seated, or lying down.

B. Positional alcohol nystagmus occurs only when the head is tipped or tilted significantly

away from “upright.”

C. Caloric nystagmus is nearly impossible to induce in a moving vehicle.

D. Rotation can induce nystagmus, but it will be inconsistent with how the officer conducts

the tests and what s/he expects to observe if intoxication were the cause.

E. Movement of objects in the visual field (e.g., heavy traffic, freight train) can cause

optokinetic nystagmus but only if the suspect pays active attention to those objects.

VII. Conclusions

A. HGN/VGN (and other SFSTs) are screening tests of divided attention that correlate to

impairment caused by intoxication.

B. SFSTs do not “prove” intoxication. However, if an officer observes performance on

SFSTs that s/he knows to be consistent with that of an intoxicated individual (from prior

training and observation of other subjects), then the officer will make the proper decision

to arrest and request a confirmatory chemical sample (breath, blood, urine, saliva).


ISSUE HIGHLIGHT

In the United States, drivers impaired* by alcohol and/ordrugs are responsible for more than 16,000 deaths, onemillion injuries, and $45 billion in costs annually.1 As

part of the attempt to reduce these human and economictolls, law enforcement officers routinely conduct tests of eyemovements to determine if a driver is under the influenceof alcohol or other drugs. Alcohol, other central nervous sys-tem (CNS)-depressant drugs, inhalants, and phencyclidine(PCP) and its analogs will affect the neural centers in thebrainstem and cerebellum, which control eye movements,as well as other motor, sensory, and cognitive integrationareas of the brain. In addition, certain antihistamines havephysiologic and cognitive effects similar to CNS-depressantdrugs.

Blood alcohol concentration (BAC), also known as blood alco-hol level, is either measured directly from a blood sampleor estimated from a breath or urine sample. BAC is com-monly reported as a percentage of alcohol weight per vol-ume of blood. When impairment is due solely to theinfluence of alcohol, most states and Canadian provincesdefine the legal limit for passenger vehicle drivers as 0.08%,while some states still allow the higher limit of 0.10%.

Positive findings on the Horizontal Gaze Nystagmus(HGN) test have been shown to correlate highly with bothBAC and cognitive impairment.2 The American Optomet-ric Association has previously recognized the validity andreliability of the HGN test as used by the law enforcementcommunity.3

Nystagmus testing in intoxicated individuals

Karl Citek, O.D., Ph.D.,a Bret Ball, O.D.,a and Dale A. Rutledge, Lieutenantb

aCollege of Optometry, Pacific University, Forest Grove, Oregon and bthe Oregon State Police, Wilsonville, Oregon

Citek K, Ball B, and Rutledge DA. Nystagmus testing in intoxi-cated individuals. Optometry 2003;74:695-710.

Background: Law enforcement officers routinely conduct psy-chophysical tests to determine if an impaired driver may beintoxicated or in need of medical assistance. Testing includesassessment of eye movements, using the Horizontal Gaze Nys-tagmus (HGN) and Vertical Gaze Nystagmus (VGN) tests, whichare conducted at roadside by patrol officers. These tests pre-viously have been validated when the subject is placed in astanding posture with head upright. However, certain condi-tions require that the subject be tested while seated or supine.Under these conditions, Positional Alcohol Nystagmus (PAN)could be induced and mistaken for HGN or VGN.

Methods: The study was conducted at law enforcement train-ing academy alcohol workshops in the Pacific Northwest.Ninety-six volunteer drinkers were tested when sober andthree times after drinking alcohol by 40 volunteer officersexperienced in administering the tests. Blood alcohol con-centration (BAC) was measured objectively with a calibratedbreath analysis instrument each time a subject was tested.

Results: The number of eye movement signs observed duringthe HGN test at any posture increases with increasing BAC.The presence of VGN at any test posture occurs only in thepresence of signs of HGN and only at high levels of impair-ment. PAN was most often observed at BACs of 0.08% andhigher, but was never confused with the observation of HGNor VGN, regardless of test posture.

Conclusions: The HGN test administered in the standing,seated, and supine postures is able to discriminate impair-ment at criterion BACs of 0.08% and 0.10%. The VGN testcan identify high levels of impairment at any test posture.Therefore, these tests can be used by an officer to determineif a driver is impaired, regardless of whether the driver isstanding, seated, or supine.

Key Words: Alcohol, blood alcohol concentration (BAC), hori-zontal gaze nystagmus (HGN), impairment, law enforcement,positional alcohol nystagmus (PAN), vertical gaze nystagmus(VGN)

695

VOLUME 74/NUMBER 11/NOVEMBER 2003 OPTOMETRY

* The inability to safely operate a motor vehicle. This may be cog-nitive (e.g., abnormal perception of time and space) or physical (e.g.,restricted use of a limb or uncorrectable vision loss).


ISSUE HIGHLIGHT

Officers conduct the HGN test at roadside as partof the Standardized Field Sobriety Tests (SFSTs).4-6

The HGN test assesses lack of smooth pursuit,sustained endpoint nystagmus, and inducednystagmus prior to a lateral gaze angle of 45degrees. Recently, the Vertical Gaze Nystagmus(VGN) test has been added to the SFST trainingfor patrol officers.4 The VGN test assesses nys-tagmus induced in upgaze.

Collectively, the SFSTs are used to establish prob-able cause for arrest on a Driving Under the Influ-ence (DUI) charge and subsequent request for abreath, blood, or urine sample, in order to objec-tively measure the BAC. These tests also are con-ducted by specially trained officers as part of theDrug Recognition Expert (DRE) evaluation whenthe presence of a drug (or drugs other than or inaddition to alcohol) is suspected.7 Results of thesetests assist the officer in accurately and reliablydetermining the presence of CNS-depressantdrugs, inhalants, and PCP.8,9

The procedure of the HGN test was standardizedmore than 20 years ago by the National HighwayTraffic Safety Administration (NHTSA).10,11 In themid 1980’s, NHTSA standardized the procedureof the VGN test as part of the DRE evaluation.12

As currently taught, both procedures require thatthe suspect stands erect with feet together, handsat the sides, and head upright, facing forward.

However, there are numerous situations in whichconducting the tests in the standing posture wouldbe unsafe or impossible. The most common ofthese occurs when the suspect is significantly tallerthan the officer, such that the officer would not beable to see the standing suspect’s eyes without seri-ously compromising the officer’s safety. Adverseweather conditions can make testing at roadsidedangerous for both the officer and the suspect. Thesuspect might be disabled or otherwise unable tostand upright as instructed. Stops at sobriety check-points may require the officer to make an initialassessment of a driver who is seated behind thewheel of the vehicle. Finally, the officer may becalled to the scene of an accident in which theinjured driver already is secured to a gurney orbackboard by paramedics. In such cases, the offi-cer must be sure the impairment and eye signs arenot due to a medical emergency, such as headinjury, stroke, or seizure, or to inappropriate, orinadvertent visual or vestibular stimulation, suchas optokinetic nystagmus or positional nystagmus.

A recent study has demonstrated a high correla-tion of HGN results between standing and seatedpostures for low BACs.13 The goals of the currentstudy are to confirm the validity and reliabilityof the HGN and VGN tests in the standing pos-ture and to establish their validity and reliabilityin the seated and supine postures for BACs up toand above the legal limit for all the United Statesand Canada.

Review of impaired eye movementsThe eye movements of an impaired individual dif-fer dramatically in appearance from those of a nor-mal, sober individual and are easily observed bya trained officer, without the need for any spe-cialized or sophisticated equipment. Fine-motorcontrol of the eyes is characterized by the abilityto make smooth-pursuit movements and to prop-erly fixate stationary targets either straight aheador to the side. Virtually all normal individuals canmake smooth pursuit eye movements to track tar-gets up to 30 deg/sec, and most can track targetsat speeds up to 100 deg/sec.14 If a target moves tooquickly for the smooth pursuit system to trackaccurately, brief catch-up saccades will be inter-posed during the eye movement and the eyes willbe seen to jerk as they follow the target. Forimpaired individuals, catch-up saccades are read-ily evident for target speeds of about 30 deg/sec.At high levels of impairment, an individual caneven lose the ability to make saccades and, thus,will be able to follow a moving target only by mov-ing the entire head and/or upper body.

Fixation of a stationary target involves the sameneural centers in the brainstem and cerebellumas smooth pursuits, and may be thought of as a“zero-velocity” pursuit eye movement.14 If fixationof a peripheral target cannot be maintained cor-rectly, the eyes will drift back toward the centerand jerk quickly toward the target. The drifttoward the center represents the slow phase of theresulting nystagmus, while the jerk toward thetarget represents the fast phase. Thus, the direc-tion of the fast phase will change with the direc-tion of gaze.

Many normal, sober individuals initially willshow one or two beats of small-amplitude nys-tagmus when the eyes are moved to extreme lat-eral gaze positions.15 This is alternately termedendpoint nystagmus or nystagmus at maximum devi-ation. The nystagmus usually dissipates within 1

696

OPTOMETRY VOLUME 74/NUMBER 11/NOVEMBER 2003


ISSUE HIGHLIGHT

to 2 seconds, if gaze is maintained at such a posi-tion. On the other hand, impaired individuals typ-ically demonstrate distinct, large-amplitudenystagmus that is sustained for several secondsat these positions.10,11

Fatigue nystagmus will occur in normal, sober indi-viduals when gaze is maintained at an extremelateral position for 30 seconds or more.15 A recentstudy suggests that lack of sleep may exaggeratenormal endpoint nystagmus,16 but no other stud-ies are known to prove that sleeplessness or sys-temic fatigue affect any other eye movements.

Gaze-evoked, gaze-induced, or, simply gaze nys-tagmus is a sustained nystagmus prior to anextreme lateral gaze position. It is indicative ofneurological damage if it occurs unilaterally orasymmetrically, and of alcohol and/or drugimpairment if it is bilateral and somewhat sym-metric.17 In addition, high levels of alcohol impair-ment, or impairment with certain drugs, eitheralone or in combination with alcohol, may pro-duce sustained, large-amplitude bilateral verticalnystagmus in upgaze but not downgaze.17

Alcohol will alter the viscosity of the endolymphin the vestibular apparatus. This will affect theindividual’s sense of balance and any eye move-ments that are influenced by the vestibular sys-tem.18 Depending on the relative concentrationsof alcohol in the blood and endolymph, positionalalcohol nystagmus (PAN) may be induced in pri-

mary gaze when the head is tipped or tilted to anon-upright position. PAN originally was con-sidered to be a very sensitive diagnostic assess-ment of alcohol intoxication.19 This may be truein a clinical or laboratory setting, but it is nothelpful to the officer in the field who does nothave the testing equipment necessary to make thecareful measurements. Nonetheless, officersmust be aware that an unintentional head tilt bythe subject may induce PAN, which may con-found or exacerbate the other eye movements theofficer is testing.

MethodsAlcohol workshopsAlcohol workshops are used to train recruits onthe use of SFSTs and to re-acquaint officers whoare training to become DREs with specifics of theSFSTs. Workshops usually last about 3 to 4 hours,during which subjects receive measured doses ofalcoholic beverages for about 2 hours, as well assnack foods. Some subjects are purposelyrecruited as “placebo drinkers,” maintaining zeroor low BACs throughout the workshop. Each sub-ject’s BAC is carefully monitored throughout theworkshop.

The current study was conducted at nine regu-larly scheduled workshops in Oregon, Washing-ton, and Idaho. Evaluations were performed byexperienced officers in a room or area separatefrom the training area, in order to avoid dis-

697


Table 1. Demographic data for the drinking volunteers in the studyFemale Male

Age (yrs) Number 37 59Mean (SD) 30.0 (8.4) 28.3 (8.1)Minimum 21.0 21.2Maximum 51.2 62.8

Weight (lbs) Mean (SD) 150.2 (36.6) 198.0 (28.6)Minimum 100 148Maximum 270 283

Ethnicity Asian 1 0African–American 0 2White 36 57

Prescription for driving Spectacles 7 8Contact lenses 9 8

Pre-test Equal pupil sizes 37 58Equal tracking 37 59

SD, Standard deviation.


ISSUE HIGHLIGHT

rupting the trainees. Subjectswere evaluated at four differenttimes during each workshop.Baseline evaluations were per-formed at the beginning of theworkshop, before the subject’sfirst drink; BAC measurementsconfirmed that all subjectsstarted with blood alcohol levelsof 0.00%. The first set of evalu-ations was conducted about 1hour after the start of drinking,the second set was conducted atthe end of the 2-hour drinkingperiod, and the final set was con-ducted at the end of the work-shop, at least 1 hour after the lastdrink. Subjects did not consumeany alcohol during the evalua-tions or BAC measurements.Subjects worked with thetrainees as part of the regularworkshop in the period betweenthe second and final sets of eval-uations.

SubjectsNinety-six volunteer drinkers—37 female and 59male—participated in the study. Subjects wererecruited from local colleges, military bases, pros-ecutors’ and attorneys’ offices, and police acad-emy offices. Each subject signed an informedconsent form.

Subjects were recruited solely on the basis of theiravailability, and not on their age, gender, weight,or ethnicity. Subject demographic data are sum-marized in Table 1. Table 1 also summarizes thetypes of prescription lenses used for driving, aswell as equality of pupil sizes and ability to fol-low a stimulus (see Test Procedures) before theconsumption of alcohol.

All subjects were of legal drinking age andacknowledged varying levels of experience withdrinking alcohol. None of the subjects reportedfatigue, presence of any health conditions, or useof any medications that precluded participationin the study. Three subjects at two workshopswere unable to complete the testing; nonetheless,their data for the portions completed areincluded in the analyses we discuss here.

EvaluatorsForty law enforcement officers, all certified DREsand/or SFST instructors, volunteered as evaluatorsfor the study. Officers had no other training dutiesor responsibilities during the workshops. Officerswere recruited solely on the basis of their avail-ability, and not on their experience or agency affil-iation. Table 2 lists the officers, their agencies, andtheir relevant experience. Several officers, notindicated in Table 2, participated in more thanone workshop each.

Each evaluator tested subjects only in one of threetest postures (see later discussion). In order tomask evaluators from the results at the differentpostures, evaluators were discouraged from dis-cussing their results during the workshop. Eval-uators also were masked from the BACmeasurements taken during the workshop.

Six evaluators were available at each workshopconducted in Washington and Idaho, and at twoof the workshops in Oregon, evaluating a total of25 female and 43 male subjects. Thus, each sub-ject was tested separately by two evaluators ateach posture at each test time. Three evaluatorswere available at each of the three remainingworkshops in Oregon, evaluating a total of 12female and 16 male subjects. These subjects were

698


Demonstration of Horizontal Gaze Nystagmus test in seated posture.Figure 1


ISSUE HIGHLIGHT

tested once at each posture at each test time.Combining data from all workshops, there was amaximum of 164 evaluations at each posture ateach test time.

Test posturesTesting was conducted on each subject in threepostures: standing, seated, and supine. The stand-ing posture was consistent with that recom-mended by NHTSA guidelines and previousvalidation studies,4,20-22 in that subjects stood withfeet together, hands at the sides, and head upright,facing forward.

In the seated posture, the subjectsat in an armless chair or foldingchair, with head upright andturned approximately 45 degreesto the side. The evaluator stoodto the same side as the subject’sturned head, such that the sub-ject always directly faced theevaluator (see Figure 1).

In the supine posture, subjectslaid flat on their back atopstacked gym mats at a height ofabout 18 inches (46 cm). Subjectswere instructed to keep theirheads straight and in line withtheir bodies for all testing, exceptPAN (see below for clarification),and evaluators were instructed toperform the tests from directlyabove the subjects (see Figure 2).

BAC measurementsBlood alcohol levels wereassessed at each test time duringeach workshop using calibratedbreath analysis instruments andprocedures equivalent to thoserequired by each state for theassessment of an actual DUI sus-pect. Certified breath analysisspecialists performed measure-ments using Intoxilyzer 5000instruments in Oregon andIdaho, and DataMaster instru-ments in Washington. To estab-lish BAC, Oregon requires onlya single reading, whereas Idahoand Washington require two

readings. All Idaho and Washington measure-ments reported here are the averages of therespective readings for each subject. The meandifference and standard deviation for all pairs ofreadings from Idaho and Washington are both0.003%.

One subject at a Washington workshop, who didnot complete the testing, was given a single meas-urement at the first and only evaluation time witha calibrated portable breath test instrument, so asto avoid possible contamination of the DataMasterinstrument.

699


Demonstration of Horizontal Gaze Nystagmus test in supine posture.Figure 2


ISSUE HIGHLIGHT

Test proceduresPre-test

At the start of the eye movement tests of theSFSTs, officers check for the presence of eye-

glasses or contact lenses, and for ocular rednessand excessive tearing.4 They also assess the sub-ject’s pupil sizes and tracking ability. Previouslyundiagnosed anisocoria may indicate a recenthead injury, such as trauma or stroke. Inability

700


Table 2. Officers, listed alphabetically by state, who volunteered as evaluators for this study, including years of experience as certified Drug Recognition Experts (DREs) and/or Standardized Field Sobriety Test (SFST) instructors

State Evaluator Agency DRE SFST

Oregon Deputy Scott Bressler Benton County SO 3 9Officer David Driscoll Salem PD 6 13Trooper Timothy Fox SP 4 6Deputy Dustin Frenzel Linn County SO — 3Officer Robert Hayes Albany PD 3 —Sergeant Lance Inman Keizer PD 3 4Trooper Michael Iwai SP 2 —Trooper Eric Judah SP 3 —Officer Kristina Knox Salem PD 5 4Officer David Leday Keizer PD 4 3Officer Tim Lenihan* Myrtle Creek PD 5 4Deputy Timothy McCall Harney County SO 1 2Trooper David Peterson SP 5 —Sergeant Robert Ruark Polk County SO 1 <1Lieutenant Trace Schreiner DPSST 3 4Officer Justin Stevenson Dallas PD 1 1Officer K.T. Taylor Sandy PD — 2Sergeant Tim Weaver* Newberg PD 2 2Trooper Steve Webster SP 6 6

Washington Trooper James Aye SP 2 —Trooper Curt Boyle SP 3 3Trooper Nathan Elias SP 1 —Trooper Steve Gardner SP 2 —Trooper Darrell Hash SP 4 3Officer Michael Henry Puyallup PD 4 3Trooper Harlan Jackson SP 3 <1Officer Theresa Kubala Vancouver PD 1 <1Trooper Bruce Lantz SP 4 <1Trooper Darrin Latimer SP 3 —Trooper Brian Mihelich Sp <1 —Trooper Shane Nelson SP 1 —Trooper D.A. O’Neill SP 4 1Officer Kelly Parsons Walla Walla PD 3 <1Deputy J. Sousley Pierce County SO 2 1Trooper Keith Trowbridge SP 3 —Trooper David Wilbur SP 3 3

Idaho Corporal Craig Boll SP 3 —Trooper T.J. Harms SP 1 —Trooper Timothy Horn SP 2 —Sergeant Timothy Johnson SP 5 3Trooper Edward Robertson Sp 2 —Corporal Lance Rogers McCall PD 1 1

SP, State Police/Patrol; PD, Police Department; SO, Sheriff’s Office/Department; and DPSST, Department of Public Safety and Training.* Participated in pilot study only.


ISSUE HIGHLIGHT

to follow the stimulus or non-congenital nystag-mus—especially in primary gaze—also may indi-cate a head injury or the presence of drugs otherthan alcohol. The appearance of “bloodshot,watery eyes” may suggest recent exposure of thesubject to a noxious environment, such as asmoke-filled room, but also occurs in response tothe dehydrating effects of ingested alcohol.

Testing normally is not performed if the subjecthas congenital nystagmus, restricted eye move-ments (i.e., noted by the officer as “inability to fol-low the stimulus”), or blindness or loss of one eye.Otherwise, spectacles are removed during testingto allow the officer to see the subject’s eyes whenthe stimulus is moved to lateral and upgaze posi-tions. The officer confirms that the subject can seethe stimulus—usually a pen, penlight, or finger-tip—before starting the test. Soft or rigid contactlenses are kept in place, as they should not affectthe testing. If they are properly fit and main-tained, they should not be displaced or fall outduring testing.

We have found that uncorrected high refractiveerror (>± 8.00 D), astigmatism (0.50 and aboveat any axis), anisometropia (more than 1 D),amblyopia (two lines difference), and strabismusare not automatic disqualifiers for conducting thetests, since the stimulus does not have a highvisual acuity demand, and since eye movementsare not necessarily restricted with these condi-tions. Other pathological conditions, in theabsence of medications that fall into any of thedrug categories described earlier, do not produceeye movements that are similar to those observedwith intoxication. For example, acquired nystag-mus in vestibular diseases,17 multiple sclerosis,23

and a rare case of glaucoma24 occur in primarygaze or with non-upright head positions. Likewise,changes in saccades and smooth pursuits with dia-betes,25 glaucoma,26 multiple sclerosis,27 and opticneuritis27 will appear different than thoseassessed during the HGN test. Viral infections,such as cold and flu, will affect eye movementsonly if there is active involvement of thevestibular system or in the presence of impairingdrugs.28

Horizontal gaze nystagmus (HGN)Testing was conducted in the same manner in alltest postures, consistent with NHTSA guide-lines.4,20 The stimulus was held in front of the

subject’s face, approximately 12 to 15 inches (30to 38 cm) from the subject’s nose and slightlyabove eye level. This elevated eye position raisedthe upper lids and allowed the evaluator a betterview of the eyes, but did not affect the results ofthe test. The subject was instructed to keep hisor her head still and follow the stimulus with theeyes only. The subject’s left eye was observed firstduring each of the three component tests.

Smooth pursuit was assessed by moving the stim-ulus to extreme left gaze and then to extreme rightgaze at about 30 deg/sec. The test was repeatedat least once for each eye. Nystagmus at maxi-mum deviation was assessed by moving the stim-ulus first to extreme left gaze, then to extremeright gaze, such that no temporal sclera showedat either position, and held at each position forat least 4 seconds. Onset of gaze nystagmus wasassessed by moving the stimulus at about 15deg/sec to each side until nystagmus wasobserved. If nystagmus was present, the evalua-tor determined whether the angle of onset wasless than 45 degrees.

The HGN test is scored by the number of signspresent for the two eyes, scoring one sign eachper eye for lack of smooth pursuit, sustained nys-tagmus at maximum deviation, and onset of gazenystagmus prior to 45 degrees. Therefore, themaximum number of signs is six. Previous lab-oratory and field validation studies have consis-tently demonstrated that the presence of four ormore signs is highly correlated with BAC at either0.10%10,11 or 0.08%.21,22,29

Vertical gaze nystagmus (VGN)Testing was conducted in the same manner in alltest postures, consistent with NHTSA guide-lines.4,20 The stimulus was held in front of thesubject’s face, approximately 12 to 15 inches (30to 38 cm) from the subject’s nose. The subject wasinstructed to keep his or her head still and followthe stimulus with eyes only. The stimulus wasraised until the subject’s eyes were in extremeupgaze, and held at that position for approxi-mately 4 seconds. Sustained vertical nystagmusindicated a positive result.

Positional alcohol nystagmus (PAN)Officers normally do not assess PAN, but it ismentioned in the training manual as a type of nys-tagmus of which they must be aware.4 PAN may

701



ISSUE HIGHLIGHT

be induced in an alcohol-impaired individual when thehead is tilted with respect tostraight ahead, with the nystag-mus present in primary gaze.Previous research has demon-strated that PAN is not inducedin a supine posture, when thehead is in line with the body.18

The presence of PAN is easilydifferentiated from the types ofnystagmus expected during theHGN and VGN tests due to thenon-upright head position andstraight-ahead gaze.

In this study, in the standing andseated postures, the presence ofPAN was assessed by having thesubject tilt the head towardeither shoulder (see Figure 3, A).In the supine posture, the subjectturned the head to the sidetoward the evaluator (see Figure3, B). In all test postures, the sub-ject maintained fixation on thestimulus held approximately 12to 15 inches (30 to 38 cm) fromthe nose. Nystagmus in primarygaze indicated a positive result.

ResultsDemographic dataThe average age of all subjectswas 29.0 years; ranging from 21to 62 years (see Table 1). Therewas no significant difference insubject ages based on gender (p = 0.351). There was a significant difference in subject weights based on gender (p = 0), with males consistently heavier than females.

The high percentage of white subjects (97%)reflects the population of the Pacific Northwest.Follow-up studies with more ethnically diversepopulations are encouraged. Thirty-two subjects(33%) wore or reported the need to wear eitherspectacles or contact lenses for driving. Lens pre-scriptions were not considered in this study, asthe only criterion was the ability of the subjectto see and follow the stimulus used by the eval-uator; no subjects had difficulty with these tasks

under the given conditions. Anisocoria was notedin a single subject in Oregon. The condition wasdetermined to be longstanding, the subject wasaware of it, and it did not affect testing in anyway. All subsequent results are reported withoutregard to gender, weight, ethnicity, or type of oph-thalmic prescription.

Blood alcohol levelsBAC measurements were taken toward the endor after each set of evaluations, on averagebetween 4.5 and 23.5 minutes from the midpointof any given set of evaluations. The longest time

702


Demonstration of test for Positional Alcohol Nystagmus (PAN) in A, standing and B, supine postures. Test for PAN in seated posture (not shown) incorporates head tiltidentical to that in standing posture.

Figure 3

A

B


ISSUE HIGHLIGHT

difference for an individual subject was about 50minutes. Since the typical elimination rate of alco-hol is about 0.015% per hour for an averageadult,30 the measurements provide an accurateassessment of the BACs of the subjects duringeach set of evaluations.

Of the 284 total measures, 156 (54.9%) were at0.08% and higher; 95 of all measures (33.5%)were at 0.10% and higher. The highest individ-ual BACs were 0.189% for a subject in Wash-ington, 0.179% for an Idaho subject, and 0.176%for an Oregon subject.

HGNBecause of variations in physiology and neu-rology in otherwise normal, sober subjects, anofficer may observe individual signs duringHGN testing that appear similar to the signsobserved when the subject is impaired.6Nonetheless, the overall number and pattern ofsigns observed in a sober subject will be dif-ferent than those seen in an impaired subject.Also, as borne out by the results of this study,

signs typically appear in the order of per-formance of the HGN test, and symmetricallyin the two eyes, with increasing levels ofimpairment.

Baseline evaluations of sober subjectsOf the 164 evaluations conducted at each test pos-ture at baseline, fewer than 10% at any posturedemonstrated (at most) three HGN signs. Thereare no significant differences based on test pos-ture for lack of smooth pursuit, sustained nys-tagmus at maximum deviation, and onset of gazenystagmus prior to 45 degrees (all p > 0.09).

Only one evaluator observed four signs (endpointnystagmus in both eyes and gaze nystagmus priorto 45 degrees in both eyes) on a single subject inthe standing posture, but these signs were notobserved by evaluators in the seated and supinepostures. At no posture during the baseline eval-uations were five or six signs observed on anysubject. There is no significant difference basedon test posture for the number of HGN signsobserved (p = 0.518).

703


For each test posture and BAC range, average number of HGN signs observed at each BAC range, with standard error bars.Figure 4


ISSUE HIGHLIGHT

Test evaluations—Analysis by BACFigure 4 shows the average number of HGNsigns, and standard error of the mean, at each testposture and range of BACs. Note that all but thelast of the non-zero BAC ranges are in incrementsof 0.02%; only one subject achieved a blood alco-hol level above 0.18% for a single measurement.

Chi-square analysis shows that there is a statis-tically significant difference in the number ofHGN signs observed based on test posture, χ2(12)= 45.49; p = 0. Compared to the standing pos-ture, evaluators typically observed fewer signs inthe seated posture and more signs in the supineposture. However, for subjects with BACs above0.06%, the greatest difference in the mean num-ber of signs observed at the different test posturesis less than one. Thus, while the differences maybe statistically significant, they are not of prac-tical significance for the officer in the field. Notethat, on average, evaluators consistently observedmore than four signs for BACs of 0.10% andhigher, and about four signs for BACs between0.08 and 0.10%.

The relationship between each subject’s BAC andthe number of HGN signs observed by each eval-uator is given by the correlation coefficient. Thecoefficients for the current study are all very highand statistically significant at p = 0: for the stand-ing posture, r = 0.63; for the seated posture, r =0.59; and for the supine posture, r = 0.59. Bycomparison, Stuster and Burns22 reported a cor-

relation coefficient of 0.65 between BAC andHGN tested in the standing posture, while McKnight et al.29 reported correlations of 0.56 and0.55 for the standing and seated postures,respectively. The correlations of the current studyare not significantly different from those reportedby either Stuster and Burns or McKnight et al. (allp > 0.13).

Nonetheless, the purpose of a sobriety test is notto estimate an individual’s BAC, but to determineif that individual is impaired: if the impairmentis due solely to alcohol intoxication, the sobrietytest can discriminate whether the individual isover or under the legal limit for BAC.10 Signaldetection theory provides several measures thatdescribe the ability of a test to discriminate at agiven criterion level.31 Sensitivity, also known asthe true positive ratio, is the proportion of sub-jects who show a positive test result to all sub-jects who actually have the given condition. Thefalse alarm rate is the proportion of subjects whoshow a positive test result to all subjects who donot have the given condition. Accuracy is the per-centage of subjects correctly identified as havingthe condition and not having the condition. Anideal test will have sensitivity equal to one, falsealarm rate of zero, and accuracy of 100%. Thedetectability index, d’, is a measure of the abilityof the test to discriminate signal from noise, or—in the present context—to determine if a test candiscriminate a finding from a random or chanceresult.32

704


Table 3. Sensitivity, false alarm rate, accuracy, and detectability index for HGN data at each test posture at two criterion blood alcohol concentration (0.08% and 0.10%)*

Posture

BAC = 0.08% Standing Seated Supine McKnight et al.

Sensitivity 0.890 0.799 0.891 0.75False alarm rate 0.367 0.285 0.462 0.32Accuracy 77.3% 76.1% 73.0% 71%d’ 1.568 1.407 1.326 1.15

BAC = 0.10% Standing Seated Supine Good and Augsburger

Sensitivity 0.956 0.887 0.969 0.96False alarm rate 0.503 0.408 0.561 0.82Accuracy 64.7% 68.9% 61.3% 90%d’ 1.698 1.442 1.708 0.88

d’, Detectability index; HGN, horizontal gaze nystagmus; and BAC, blood alcohol concentration.* Included for comparison are calculations based on the data recorded by McKnight et al.,29 testing in a seated posture, and Good and Augsburger,

5 testing in a standing posture.


ISSUE HIGHLIGHT

Sensitivity, false alarm rate, accuracy, and d’ ofthe HGN test at each test posture for each of twocriterion BACs, 0.08% and 0.10%, are shown inTable 3. Sensitivity and accuracy are consistentlyvery good for all measures, and false alarm ratesare acceptable, given the fact that the result of theHGN test provides only one of many possiblepieces of evidence of impairment to an officer.4All d’s are significant at p = 0, indicating that theevaluators could correctly discriminate impair-ment at all postures for either criterion BAC. Table3 also shows comparable results for Good andAugsburger5 and McKnight et al.29 Note that

Good and Augsburger used a criterion BAC of0.10%, which was the legal limit in Ohio at thetime of that study, with HGN conducted in thestanding posture. In contrast, McKnight et al. useda criterion BAC of 0.08%, with HGN conductedin a seated posture.

Inter-posture and inter-evaluator reliabilitySeparate evaluators conducted testing at the dif-ferent postures. In addition, most tests were con-ducted by two evaluators at each posture,allowing an assessment of test–retest reliability.

705


Table 4. Inter-posture reliabilities, test–retest reliabilities, and test–retest accuracy for the HGN test conducted at different postures and by pairs of evaluators*

Reliability Accuracy

Posture: Seated Supine Test–retest Test–retest

Standing 0.672 0.616 0.589 76.1%Seated — 0.638 0.653 73.1%Supine — — 0.713 84.7%

HGN, Horizontal gaze nystagmus.* All reliabilities are significant at p = 0.

Percentage of evaluations at each test posture in which VGN was observed for the given BAC ranges.Figure 5


ISSUE HIGHLIGHT

The reliability of the HGN test, both between dif-ferent postures and for test–retest, is given by thecorrelation coefficient. For psychomotor tests,such as HGN, a highly reliable test has a corre-lation coefficient of about 0.7.33 Test–retest accu-racy is a measure of the consistency betweenevaluators.

Table 4 shows the inter-posture and test–retestreliabilities, as well as the test–retest accuraciesfor the current study. By comparison, the HGNtest conducted in the standing posture previ-ously has been shown to have test–retest reli-ability of 0.59:11 this is not significantlydifferent from any of the test–retest reliabilitiesfor the current study (all p > 0.40). On the otherhand, McKnight et al.13 reported a correlationof 0.94 between the results of the standing andseated postures. This is very likely due to thefact that the same evaluator tested each subjectin both postures in that study, whereas the cur-rent study used different evaluators for eachposture.

For all test postures, there is no correlationbetween evaluator experience and the number ofHGN signs observed. All correlation coefficientsare close to zero (|r| < 0.15), and not significant(all p > 0.05).

VGNVGN is not expected in normal, sober subjects inthe absence of neurological problems. With theuse of alcohol alone, VGN may not appear untila high level of impairment is achieved, as definedfor the individual subject.4 VGN may be presentwhen other CNS-depressant drugs, inhalants, orPCP are used, either separately or in combination,or with alcohol.

Baseline evaluations of sober subjectsOf the 164 evaluations conducted at each test pos-ture at baseline, VGN was observed on only a sin-gle subject by one evaluator in the supine posture.However, VGN was not observed on the samesubject by the same evaluator at the first evalu-

706


Table 5. Sensitivity, false alarm rate, accuracy, and detectability index for VGN data at each test posture at two criterion blood alcohol concentrations (0.08% and 0.10%)*

Posture

BAC = 0.08% Standing Seated Supine

Sensitivity 0.215 0.289 0.485False alarm rate 0.032 0.041 0.072Accuracy 0.560 0.594 0.687d’ 1.066 1.184 1.420

BAC = 0.10% Standing Seated Supine

Sensitivity 0.268 0.371 0.610False alarm rate 0.065 0.080 0.144Accuracy 0.718 0.739 0.775d’ 1.897 1.074 1.341

d’, Detectability index; VGN, vertical gaze nystagmus; and BAC, blood alcohol concentration.

Table 6. Inter-posture reliabilities, test–retest reliabilities, and test–retest accuracy for the VGN test conducted at different postures and by pairs of evaluators*

Reliability Accuracy

Posture: Seated Supine Test–retest Test–retest

Standing 0.324 0.357 0.369 85.4%Seated — 0.391 0.401 83.0%Supine — — 0.515 79.7%

VGN, Vertical gaze nystagmus.* All reliabilities are significant at p = 0.


ISSUE HIGHLIGHT

ation, when the subject had a blood alcohol levelof 0.02%. It also was not observed by any otherevaluator in any other test posture either at base-line or the first evaluation.

Test evaluations—Analysis by BACFigure 5 shows the percentage of evaluations atwhich VGN was observed at each test posture forthe given BAC ranges. Chi-square analysis showsthat there is a statistically significant differencein the observation of VGN based on test posture,χ2(2) = 44.43; p = 0. Compared to the standingposture, VGN typically was observed more fre-quently in the seated and supine postures.

However, the differences based on test posture areonly evident for BACs at 0.08% and higher. Of the221 evaluations conducted at each test posture onsubjects with BACs below 0.08%, VGN wasobserved only on seven subjects (3.2%) in thestanding posture, nine subjects (4.1%) in theseated posture, and 16 subjects (7.2%) in thesupine posture. These findings do not differ sig-nificantly (p = 0.112). On the other hand, for sub-jects with BACs of 0.08% and higher, VGN wasobserved in 21.5% of evaluations in the standingposture, 28.9% in the seated posture, and 48.5%in the supine posture. At BACs of 0.10% andhigher, the percentages of observations at eachposture were 26.8%, 37.1%, and 61.0%, respec-tively.

The correlation coefficients, relating each subject’sBAC to the observation of VGN by each evalua-tor, are all good and statistically significant (all p = 0): for the standing posture, r = 0.35; for the seated posture, r = 0.37; and for the supineposture, r = 0.52.

Sensitivity, false alarm rate, accuracy, and d’ ofthe VGN test at each test posture for each of twocriterion BACs, 0.08% and 0.10%, are shown in

Table 5. All d’s are significant at p = 0. While thesensitivities are all relatively low, the false alarmrates are excellent, and the accuracies are verygood.

Inter-posture and inter-evaluator reliabilityReliabilities between test postures and pairs ofevaluators, and test–retest accuracies, weredetermined as for the HGN test discussed earlier.Table 6 shows the inter-posture and test–retestreliabilities, as well as the test–retest accuraciesfor the VGN test.

For all test postures, there is no correlationbetween evaluator experience and the obser-vation of VGN. All correlation coefficients areclose to zero (|r| < 0.09), and not significant (allp > 0.05).

Combined results of HGN and VGN testsTable 7 shows the number of evaluations at eachtest posture in which HGN and VGN wereobserved. Presence of HGN is determined by theobservation of at least four signs during testing.The data are consistent with the fact that whenimpairment is due to alcohol and/or drugs, VGNwill be present only when HGN is present.4 Forthe single evaluation in the standing posture inwhich VGN was observed and HGN was not, theevaluator did note two signs of HGN and thesubject’s BAC was 0.09%. All six evaluations inthe seated posture, in which VGN was observedand HGN was not, were conducted by the sameevaluator. The evaluator noted either two orthree HGN signs for each evaluation, and threeof these subjects had BACs above 0.08%. Thus,the observation of VGN alone would have cor-rectly identified four of the seven subjects asabove the 0.08% limit for BAC. In all sevencases, it is most likely that the evaluators expe-rienced difficulty observing the subjects’ eyes,

707


Table 7. Number of evaluations at each test posture in which HGN and VGN were observed*Posture

HGN observed? VGN observed? Standing Seated Supine

Yes Yes 62 79 144Yes No 251 194 193No Yes 1 6 0

HGN, Horizontal gaze nystagmus and VGN, vertical gaze nystagmus.* Presence of HGN is determined by the observation of at least four signs during testing.


ISSUE HIGHLIGHT

but the data demonstrate that this was not awidespread or overwhelming problem for thestudy.

PANResults are presented to demonstrate that officerscan correctly identify and distinguish PAN fromother types of nystagmus. It is not the intentionof this study to include the observation of PANduring an actual DUI or DRE evaluation. Thus,it is of little value to report sensitivity analysis val-ues or reliabilities. For the interested reader, thosevalues are very similar to those reported for theVGN test discussed earlier.

Baseline evaluations of sober subjectsOf the 164 evaluations conducted at each test pos-ture at baseline, PAN was observed on only twosubjects at one workshop by the same evaluatorin the standing posture. However, PAN was notobserved on the same subjects by any of the otherfive evaluators.

Test evaluations—Analysis by BACChi-square analysis shows that there is a statis-tically significant difference in the observation ofPAN based on test posture, χ2(2) = 41.80; p = 0.PAN was observed with approximately equal fre-quency in the standing and seated postures, butwith greater frequency in the supine posture.Nonetheless, PAN was observed in fewer than10% of all evaluations with BAC below 0.08%. Inaddition, because of the head tilt required toinduce PAN, it should never be mistaken as a signof HGN or VGN.

DiscussionConsistent with previously published results, weconfirm the validity of the HGN test in the stand-ing posture to discriminate blood alcohol levelsof 0.08% and 0.10%. We also establish, with sim-ilar accuracies and reliabilities, the use of theHGN test in the seated and supine postures. Theaverage inter-evaluator reliability and accuracydemonstrate that HGN is a highly reliable test.

However, there were statistically significant dif-ferences in the observation of HGN based on testposture. We attribute these differences to the abil-ity of the evaluator to detect the signs, rather thanto incorrectly identify PAN as a sign of HGN.

Evaluators conducting the test in the seated pos-ture occasionally reported difficulty seeing thesubject’s eye that was opposite the head turn. Onthe other hand, evaluators conducting the test inthe supine posture could easily shift positioneither along or across the subject’s body to bet-ter observe the eyes during each part of the test.

Nonetheless, these differences do not suggest thatimpaired seated subjects would be mistaken assober, nor that sober supine subjects would bemistaken as impaired. As shown in Figure 4, eval-uators typically observed fewer than two signs onsubjects with BACs below 0.04%, and four ormore signs on subjects with BACs at 0.10% andhigher, regardless of posture. For subjects withBACs between 0.08% and 0.10%, evaluatorsobserved (on average) about 4.5 signs in the stand-ing and supine postures and 3.9 signs in theseated posture. While statistically significant,these differences are not of practical significanceto the officer in the field.

We recommend that the officer who needs to con-duct the HGN test in the seated posture positionthe subject in such a way that the subject’s eyescan be seen easily throughout the test. This mayinvolve asking the subject to turn the bodyslightly at the waist, in addition to the head turnused in the current study. Such a minor changein posture will not affect the results.

We also confirm that VGN is present only whensigns of HGN are present, and that the VGN testcan be used to identify high levels of impairmentat any test posture. Again, we attribute the sta-tistical difference in observation of VGN at the dif-ferent postures to the ease with which theevaluators could detect the nystagmus, rather thanthe influence of the postures themselves. Asshown in Figure 5, fewer than 10% of subjectswith BACs below 0.08% exhibited VGN at anyposture, whereas at least 30% of subjects withBAC at and above 0.12% exhibited VGN.

ConclusionOfficers in the field observe various indicators ofa driver’s impairment, including driving behav-ior, physical signs, and performance on psy-chophysical tests. We conclude that the proper useof the HGN and VGN tests at any test posture willhelp an officer correctly identify individuals

708



ISSUE HIGHLIGHT

impaired with alcohol at BACs of 0.08% andhigher. By extension, since other CNS depressantdrugs, inhalants, and PCP affect the same neuralcenters as alcohol, DRE officers can use the sametests and test postures to aid in identification ofimpairment with substances other than, or inaddition to, alcohol.

AcknowledgmentsWe would like to thank the training coordinators in each state for allowing usto conduct this study at their alcohol workshops and for encouraging officers toparticipate: Captain Charles Hayes, Oregon State Police; Sergeant Steve John-son, Washington State Patrol; and Trooper Tim Riha, Idaho State Patrol. We alsowish to thank the training specialists and technicians who operated the breathanalysis instruments and helped coordinate the workshops: Mr. Hal Merrill,Agency Trainer, and Ms. Kathy Irwin, Training Coordinator, Oregon Departmentof Public Safety Standards and Training; Mr. William Bogen, Breath Alcohol Sec-tion, and Troopers B.I. Bowers and Debbie Laur, Washington State Patrol; andSergeant Dean Matlock, Idaho State Patrol. We very much appreciate the will-ingness of the subjects to take part in the many additional evaluations comparedto the standard alcohol workshop protocol. We are grateful to Detective MikeHerb, Forest Grove (Oregon) Police Department, and Messrs. James Bewley, DustyBodman, Levi Porter, and Michael Johnson for the demonstration photographsof the test procedures. We are especially indebted to Dr. Jack Richman for hisinsightful comments and suggestions in reviewing drafts of this manuscript. Lastly,we would like to express our sincere gratitude to the officers who volunteeredtheir time and expertise as evaluators. The opinions expressed in this report aresolely those of the authors, and do not necessarily reflect those of the officersnamed in the report or the individuals acknowledged, nor those of the agenciesor institutions by whom they are employed. Also, they are not presented to rep-resent any position or endorsement held or taken by the American OptometricAssociation.

References1. Michie J. First-time study highlights need for increased

awareness of substance use and driving behaviors.http://www.health.org/pressrel/dec98/10.htm, Last accessedDecember 26, 1998.

2. Kennedy RS, Turnage JT, Rugotzke GG, et al. Indexing cog-nitive tests to alcohol dosage and comparison to stan-dardized field sobriety tests. J Stud Alcohol 1994;55:615-28.

3. Resolution 6: Horizontal gaze nystagmus as a field sobri-ety test. J AM OPTOM ASSOC 1993;64:653.

4. National Highway Transportation Safety Administration.DWI Detection and Standardized Field Sobriety TestingStudent Manual, 2002. Publication HS 178 R1/02.

5. Good GW, Augsburger AR. Use of horizontal gaze nys-tagmus as a part of roadside sobriety testing. Am J OptomPhysiol Opt 1986;63:467-71.

6. Halperin E, Yolton RL. Is the driver drunk? Oculomotorsobriety testing. J AM OPTOM ASSOC 1986;57:654-57.

7. Kosnoski EM, Yolton RL, Citek K, et al. The drug eval-uation classification program: using ocular and other signsto detect drug intoxication. J AM OPTOM ASSOC 1998;69:211-27.

8. Adler EV, Burns M. Drug recognition expert (DRE) val-idation study. Arizona Governor’s Office of HighwaySafety, 1994.

9. Smith JA, Hayes CE, Yolton RL, et al. Drug recognitionexpert evaluations made using limited data. Forensic SciInt 2002;130:167-73.

10. Burns M, Moskowitz H. Psychophysical tests for DWIarrest. DOT HS-802-424. National Technical InformationService, Springfield, Va., 1977.

11. Tharp V, Burns M, Moskowitz H. Development and fieldtest of psychophysical tests for DWI arrest. DOT HS-805-864. National Technical Information Service, Springfield,Va., 1981.

12. Bigelow GE, Bickel WE, Roache JD, et al. Identifyingtypes of drug intoxication: laboratory evaluation of a sub-ject-examination procedure. DOT HS-806-753. NationalTechnical Information Service, Springfield, Va., 1985.

13. McKnight AJ, Langston EA, McKnight AS, et al. Sobri-ety test for low blood alcohol concentrations. Accid AnalPrev 2002;34:305-11.

14. Robinson DA. Eye movement control in primates: theoculomotor system contains specialized subsystems foracquiring and tracking visual targets. Science 1968;161:1219-24.

15. Abel LA, Parker L, Daroff RB, et al. End-point nystag-mus. Invest Ophthalmol Vis Sci 1978;17:539-44.

16. Booker JL. End-position nystagmus as an indicator ofethanol intoxication. Sci Justice 2001;41:113-6.

17. Dell’Osso LF, Daroff RB. Nystagmus and saccadic intru-sions and oscillations. In: Duane’s Clinical Ophthalmology1997;2(11).

18. Fregly AR, Bergstedt M, Graybiel A. Relationshipsbetween blood alcohol, positional alcohol nystagmus andpostural equilibrium. Q J Stud Alcohol 1968;28:11-21.

19. Aschan G, Bergstedt M, Goldberg L, et al. Positional nys-tagmus in man during and after alcohol intoxication. Q J Stud Alcohol 1956;17:381-405.

20. National Highway Transportation Safety Administration.Drug Evaluation and Classification Training: The DrugRecognition Expert School, 2002. Publication HS 172AR5/02.

21. Anderson TE, Schweitz RM, Snyder MB. Field evaluationof a behavioral test battery for DWI. DOT HS-806-475.National Technical Information Service, Springfield, Va.,1983.

22. Stuster J, Burns M. Validation of the standardized fieldsobriety test battery at BACs below 0.10 percent. SantaBarbara, Calif.: Anacapa Sciences, Inc., August 1998.

23. Averbuch–Heller L. Acquired nystagmus. Curr TreatOptions Neurol 1999;1:68-73.

24. Kansal S, Miller M. Bilateral Duane syndrome with bilat-eral congenital glaucoma. J AAPOS 2001;5:325-6.

25. Virtaniemi J, Laakso M, Nuutinen J, et al. Voluntary eyemovement tests in patients with insulin-dependent dia-betes mellitus. Acta Otolaryngol 1993;113:123-7.

26. Severt WL, Maddess T, Ibbotson MR. Employing fol-lowing eye movements to discriminate normal from glau-coma subjects. Clin Experiment Ophthalmol 2000;28:172-4.

27. Reulen JP, Sanders EA, Hogenhuis LA. Eye movement dis-orders in multiple sclerosis and optic neuritis. Brain1983;106(Pt 1):121-40.

28. Baker SD, Borys DJ. A possible trend suggestingincreased abuse from Coricidin exposures reported to theTexas Poison Network: comparing 1998 to 1999. Vet HumToxicol 2002;44:169-71.

709



ISSUE HIGHLIGHT

29. McKnight AJ, Langston EA, Lange JE, et al. Developmentof Standardized Field Sobriety Tests for Lower BAC Lim-its. Washington, D.C.: National Public Services ResearchInstitute, 1995.

30. Julien RM. A Primer of Drug Action. New York: WH Free-man, 2001:56-8.

31. Green DM, Swets JA. Signal Detection Theory and Psy-chophysics. New York: John Wiley and Sons, 1966.

32. Green DM. Psychoacoustics and detection theory. In:Swets JA, ed. Signal Detection and Recognition by HumanObservers: Contemporary Readings. New York: JohnWiley and Sons, 1964:70.

33. Bruning JL, Kintz BL. Computational Handbook of Sta-tistics, 3rd ed. Glenview, Ill.: Scott, Foresman, and Co.,1987:226.

Corresponding author:

Karl Citek, O.D., Ph.D.Pacific University

College of Optometry2043 College Way

Forest Grove, Oregon 97116

[email protected]

710



PAPER

GENERAL; TOXICOLOGY

Karl Citek,1 O.D., Ph.D.; Ashlee D. Elmont,2 O.D.; Christopher L. Jons,3 O.D.;Chad J. Krezelok,4 O.D.; Joseph D. Neron,5 O.D.; Timothy A. Plummer6;and Timothy Tannenbaum7

Sleep Deprivation Does Not Mimic AlcoholIntoxication on Field Sobriety Testing*

ABSTRACT: Previous research shows that sleep deprivation (SD) produces cognitive impairment similar to that caused by alcohol intoxication.Individual studies suggest that SD also causes deficits in motor skills that could be mistaken for intoxication. Consequently, SD often is used as adefense when an impaired driver is charged with driving while intoxicated. Twenty-nine adult subjects participated in two test sessions each, one aftera full night’s rest and the other after wakefulness of at least 24 h. Subjects consumed prescribed amounts of alcohol during each session. Lawenforcement officers conducted field sobriety tests identical to those with which a driver would be assessed at roadside. Researchers also measuredclinical responses of visual function and vital signs. The presence and number of validated impairment clues increase with increasing blood alcoholconcentration but not with SD. Thus, SD does not affect motor skills in a manner that would lead an officer to conclude that the suspect is intoxi-cated, unless intoxication also is present.

KEYWORDS: forensic science, sleep deprivation, fatigue, alcohol, intoxication, field sobriety test, nystagmus, blood alcohol concentra-tion, driving while intoxicated

Sleep deprivation (SD) greatly increases the risk of motor vehiclecrashes (1–6), but jurisdictions in the United States and around theworld are only recently beginning to establish legal consequences fordrowsy or sleep-deprived driving (7–9, http://www.drowsydriving.org/docs/DrowsyDrivsing Prevention Week 2008 Press Release.pdf).Many people either are unaware of the danger of driving with SDor endure it as part of their occupations (10–16). To assist driverswho may be drowsy, an in-vehicle monitoring system has beendeveloped and finally introduced on production vehicles (17, http://www.mbusa.com/mercedes/#/advancedTechOverview/press/).

SD produces impairment similar to alcohol intoxication based onassessments of cognitive and cognitive-motor function (18–25) andsimulated driving performance (26,27). SD significantly impairsattention skills in all individuals, with young drivers more affectedthan older drivers on tests of reaction time (28). Useful visual fielddecreases with age (29) and intoxication (30), and SD causes fur-ther reduction regardless of age (31). While the specific drivingskills (32–34) and physiological mechanisms (35) affected by SDand intoxication may differ, even low levels of alcohol intoxicationcombined with partial or full SD cause substantial decrements insimulated driving performance (36–40).

SD has been shown to affect saccadic eye movements (41,42),pupil size in total darkness (43), pupil reaction to light (41,44) andemotional stimuli (45,46), and blink rate (47). All of these studiesalso report subjective changes in perception of sleepiness at differ-ent levels of SD. Except for pupil reaction to light, law enforce-ment officers do not assess any of these physiological factors onsuspected impaired drivers (48,49).

Several reports suggest that SD directly produces changes invisuomotor functions that could be mistaken as being caused byintoxication (42,50,51). Additional studies suggest that SD exacer-bates or prolongs the effects of intoxication on eye movements(52–54). Consequently, SD often is offered as a defense when animpaired driver is charged with driving while intoxicated (DWI).(Note that different jurisdictions may use different, but related,legal terms, such as driving under the influence [DUI], drivingunder the influence of intoxicants [DUII] or operating [a vehicle]while intoxicated [OWI].) As most jurisdictions do not yet make ita crime to drive sleep deprived, an intoxicated driver could escapesevere legal and civil penalties if he can convince the judge andjury that he merely was tired.

Many law enforcement officers have reported to the authors thatthey can distinguish between intoxicated and sleep-deprived drivers.However, no prior research has assessed SD under the actual psy-chophysical procedures used by officers to evaluate driver impair-ment, known as field sobriety tests (FSTs). In addition, doctorsoccasionally are asked to testify about an officer’s findings, buttheir expertise may be limited to clinical testing, which often is dif-ferent in procedure (including test protocol and equipment orinstruments used), expected findings, and interpretation. The goalof this research is to determine whether SD causes changes in per-formance on FSTs and related clinical tests in a manner that couldbe confused with intoxication.

1Pacific University College of Optometry, 2043 College Way, ForestGrove, OR.

2Private Practice, Boulder, CO.3Private Practice, Buffalo, WY.4Private Practice, Bozeman, MT.5Private Practice, Beaverton, OR.6Oregon State Police, Salem, OR.7Washington County (Oregon) Sheriff’s Office, Hillsboro, OR.*Funding support for subject payments and expenses other than alcohol was

provided through a Pacific University Faculty Development Grant to KC.Received 12 Mar. 2010; and in revised form 25 July 2010; accepted 9

Aug. 2010.

J Forensic Sci, September 2011, Vol. 56, No. 5doi: 10.1111/j.1556-4029.2011.01813.x

Available online at: onlinelibrary.wiley.com

1170 � 2011 American Academy of Forensic SciencesMDJA Annual Conference | August 11-12, 2020

Note that fatigue often is used as a synonym for SD. However,it also can be defined as specific changes in physiological or cogni-tive performance, responsiveness or appearance following continuedexertion or stimulation, such as in the well-known phenomenonfatigue nystagmus (55). Because such changes typically are unre-lated to those observed with SD, and to avoid confusion, we willnot use the term fatigue in place of SD.

Methods

Subjects

Potential subjects were either students at Pacific University, For-est Grove, Oregon, or friends or spouses of students, comprising asample of convenience. Before beginning the study, candidatesreviewed and signed informed consent and model release forms,approved by the Pacific University Institutional Review Board.Candidates completed a detailed questionnaire regarding personaland health histories and experience with consuming alcohol. Thosewho admitted to a history of alcohol or substance abuse, use of cer-tain medications, pregnancy, or presence of any medical conditionwith which alcohol consumption is contraindicated were excludedfrom the study. Individuals who were excessively over- or under-weight, as determined by body mass index of >40 or <18, respec-tively, likewise were excluded for health reasons. One male subjecttaking medication for hypertension was allowed to participate sub-sequent to his doctor’s approval; the medication was not known tohave contraindications for alcohol use nor cause any side effectsthat would confound the results.

Each subject was asked to participate in two test sessions at least1 week apart, one after a full night’s rest and the other after wake-fulness of at least 24 h. Subjects were arbitrarily assigned to eachsession based on availability.

Twenty-nine subjects (14 women, 15 men) qualified for andcompleted the study. Their demographic information is provided inTable 1. All were 21 years of age or older, as confirmed by a validdriver’s license. Only one subject was over 34 years of age, but theresults of this sole 52-year-old female subject are consistent withthose of the other subjects, so there is no reason to isolate orexclude her data. Two subjects are authors of this study; theremaining 27 subjects each received a $20 gift card for theirparticipation.

Breath Analysis

A calibrated breath analysis instrument, Intoxilyzer 5000 (CMI,Owensboro, KY), identical to one used for actual DWI investigationsin Oregon was used to estimate each subject’s blood alcohol concen-tration (BAC) multiple times throughout each session, as indicatedbelow. A researcher certified under Oregon guidelines to operate theinstrument collected all the BAC data. Oregon Administrative Rulesrequire that ‘‘[t]he operator is certain that the subject has not takenanything by mouth..., vomited, or regurgitated liquid from the stom-ach into mouth, for at least 15 min before taking the test’’ (http://arcweb.sos.state.or.us/rules/OARS_200/OAR_257/257_030.html). Thestudy protocol described below allowed for testing to be com-pleted during the required 15-min ‘‘deprivation period’’ prior toeach breath test, as well as to ensure that each test was conductedin close proximity to its respective breath test.

Evaluators

Six law enforcement officers volunteered as evaluators for FSTs.All officers were certified Drug Recognition Experts (DREs) withextensive experience in identifying and assessing impaired drivers.The number of evaluators at each test session varied based on theiravailability and the number of subjects present, and most evaluatorsparticipated in multiple sessions. Researchers assessed vital signsand clinical tests as described below. Test order varied based onthe availability of subjects, evaluators, and researchers within eachtest session.

FSTs

FSTs were conducted using procedures identical to those withwhich an impaired driver would be assessed. The tests includedhorizontal gaze nystagmus (HGN), vertical gaze nystagmus (VGN),walk-and-turn (WAT), one-leg stand (OLS), Romberg balance(RB), and lack of convergence (LOC). HGN, VGN, WAT, andOLS comprise the standardized FSTs that an officer typicallywould use to assess an impaired driver at roadside (48,56); thesetests, as well as RB, LOC, vital sign evaluations (see below), andothers not evaluated in this study, are part of the drug evaluationconducted by DREs (49,57). In all but two sessions, individualevaluators assessed only either of the following test combinations:HGN ⁄ VGN ⁄ LOC or WAT ⁄ OLS ⁄ RB. In the remaining two ses-sions, each of the two evaluators at each session conducted all ofthe assessments on half of the subjects.

The HGN test is comprised of three subtests, the results ofwhich were each recorded separately for each eye tested: lack ofsmooth pursuit (LSP); distinct and sustained nystagmus at maxi-mum deviation (DSNMD); and onset of nystagmus prior to 45degrees (ONP45). Details of the test procedures, scoring, andinterpretation of the HGN and VGN tests are described elsewhere(48).

Details of the test procedures and scoring of the WAT and OLStests also are described elsewhere (48). For the OLS test, evaluatorsrecorded both the validated clues observed (OLS clues) and thenumber to which the subject counted during the 30-sec test interval(OLS count). For the RB test, evaluators determined the presenceof side-to-side, front-to-back, and circular sway (RB sway); thepresence of leg, body, and eyelid tremors (RB tremors); and theactual time it took subjects to estimate the passage of 30 sec witheyes closed (RB time). Normal range of RB time is 25–35 sec(49). To avoid practice effects, WAT, OLS, and RB were assessedonly at the beginning and end of each session; subjects specifically

TABLE 1—Demographic data for subjects who completed the study.

Female Male

Number 14 15Age, years

Mean 27.3 25.45Standard deviation 7.92 2.24Range 21–52 22–30

Weight, kgMean 59.8 83.0Standard deviation 6.61 17.67Range 50.0–75.0 61.4–136.4

Body mass indexRange 18.9–27.5 19.4–39.7

Hours awake at start of workshopNormal sleep

Mean 4.0 4.1Standard deviation 1.59 1.73Range 2.0–7.2 2.5–7.9

Sleep deprivedMean 30.3 29.4Standard deviation 0.80 1.80Range 28.8–32.1 24.4–31.9

CITEK ET AL. • SLEEP DEPRIVATION AND INTOXICATION 1171


were told not to perform or practice these tasks at times other thanwhen they were being evaluated.

When this study was conducted, the procedure for the LOC testwas to assess whether the subject could converge the eyes to astimulus brought along the midline to the bridge of the nose (58).Since then, the procedure has been changed to assess convergenceonly to within 2 in. (5 cm) from the bridge of the nose (51). Evenso, we believe that our results are valid, especially because wecompared subjects to their own performance at baseline and weconducted a related clinical test at the same time.

Vital Signs

Blood pressure (BP) was measured manually with a calibratedsphygmomanometer and stethoscope; normal ranges for systolicand diastolic BP are 120–140 and 70–90 mmHg, respectively (49).Pulse rate was measured at the radial artery for 30 sec and multi-plied by two to arrive at beats per min (bpm); normal range is60–90 bpm (49).

Pupil size varies with light level and other factors, such asconvergence and emotional arousal. Typical room light providesilluminance of 300–500 lux, and a dimmer room is expected toresult in larger pupils. Pupil size was measured for both eyes usingcalibrated cards with either circles or semi-circles in diameters from1.0 to 9.0 mm in 0.5-mm increments; normal range in room lightis 2.5–5.0 mm (49). The test facility did not have a readily accessi-ble ‘‘dark room’’ to allow us to evaluate, in a timely manner, pupilsizes and responses under the additional light conditions specifiedwithin the DRE protocol.

Clinical Tests

Researchers performed additional tests similar to the manner inwhich they are conducted clinically. Nearpoint of convergence(NPC) is related to the LOC test, except that the actual distance isrecorded when the subject loses convergence; normal breakpoint is5 cm from the bridge of the nose (59). Endpoint nystagmus (EN)involves the observation of any nystagmus at extreme lateral gazeof either eye. It is distinguished from DSNMD by the fact that ENusually is of small amplitude and possibly difficult to observe(60,61), hence not distinct, and of short duration, typically dissipat-ing after 1–2 sec (50,62), hence not sustained.

Horizontal attentional field of view (AFOV) was assessed with anarc perimeter with 30-cm radius. The subject binocularly fixated acentral spot 1.4 deg in diameter. A second target, also 1.4 deg indiameter, was introduced in the far periphery of either eye and man-ually moved toward the center at a speed of about 2 deg ⁄ sec untilthe subject first reported seeing it. Consequently, AFOV representsthe lateral peripheral awareness for the left and right eyes together.

Overnight Observation Period

All but three subjects assigned to be awake for at least 24 hprior to a test session arrived at Pacific University College ofOptometry at about 10 PM on the evening before the session, aftera full day of classes or work. Subjects stayed in the student loungearea and were allowed to study, play games, and watch moviesthroughout the night, monitored by shifts of researchers. Subjectsand monitors were provided with snack and breakfast foods, softdrinks, and water. The remaining three subjects worked in an over-night medical clerical office away from campus. These subjectsmonitored themselves for wakefulness, as well as performing theregular checks of vital signs described below.

Each subject’s vital signs—BP, pulse rate, pupil sizes in roomlight—were checked four times at regular intervals throughout theovernight period, between about 10 PM and 11 AM. Lighting inthe student lounge area was less than about 100 lux during thenight, i.e., before 7 AM, for the first three measurements. For thelast measurement after 8 AM, natural light, but no direct sunlight,entered the southward-facing windows, thereby increasing ambientlight level to about 200–300 lux. Lighting in the medical office forthe three respective subjects was similar, and their data are includedwith those of the other subjects.

Test Sessions

Alcohol dosing and testing was conducted in a training room atWashington County Sheriff’s Office in Hillsboro, Oregon. A totalof nine sessions were held either on a Friday or Saturday afternoon,commencing between about 12 Noon and 1 PM, based on avail-ability of subjects and the facility. Subjects were requested to haveat most a light breakfast no sooner than about 3 h prior to the startof a test session. Sleep-deprived subjects, except for the three medi-cal office workers, were driven from Forest Grove to Hillsboro,about 6 miles, by the researchers. Well-rested subjects, as well asthe three subjects who worked in the medical office overnight, pro-vided their own transportation to the facility. Sessions finishedbetween about 3:30 and 5 PM. All subjects had designated driversavailable to take them home after each session, regardless of theirBACs at the end of the session.

Evaluators were not told whether subjects were well rested orsleep deprived. Subjects did not interact with evaluators and research-ers other than for testing purposes. Subjects, evaluators, and research-ers (other than the lead researcher and breath test operator) weremasked to the BAC readings. Evaluators and researchers did not con-fer regarding their findings. Evaluators and researchers used separatecheck-off and fill-in datasheets for each group of subjects tested ateach test period, turning those in to the lead researcher at the end ofthe period to avoid comparison of current and previous findings.

Each subject received a prescribed total dose of 80-proof liquorof his or her choice (vodka, gin, rum, or whiskey). The dose wasbased on the subject’s gender, weight, and intended BAC after 2 hof drinking, as calculated using a formula given by Jones (63). Thedose was divided into two equal portions, usually with water addedto mask the actual amount of alcohol being served. Subjects couldadd ice and as much of any mixer they wanted (orange juice, tonicwater, cola, etc.) to each portion.

Three subjects within each state of restedness were maintainedas placebo drinkers, receiving just enough alcohol to create a breathodor of alcohol and the impression that they were drinking liketheir fellow subjects. Fellow subjects, evaluators, and researchers(other than the lead researcher and breath test operator) weremasked as to the identities of the placebo drinkers. For nonplacebodrinkers, the intended goal at the end of 2 h of drinking was BACof between 0.08 and 0.11 g ⁄ dL.

Test Periods

All vital signs and test measures were assessed at the start ofeach session (baseline), before the consumption of any alcohol.Each subject began each session with a BAC of 0.00 g ⁄ dL, con-firmed with the Intoxilyzer 5000, and with no other intoxicatingdrugs present, based either on self-report or on observation duringthe overnight period.

Subjects were requested to consume the first portion of the alco-hol dose within the first 45 min after the start of drinking. To

1172 JOURNAL OF FORENSIC SCIENCES


maximize alcohol absorption, subjects were not allowed to eat anyfood during this time. All vital signs and test measures, other thanWAT, OLS, and RB, were assessed after the first portion of alco-hol was consumed (first period). Thus, subjects’ BACs were mea-sured about 1 h after starting drinking.

Subjects were requested to consume the second portion ofthe alcohol dose within the first 45 min of the second hour of thesession. Subjects now were allowed to eat snack foods during thisperiod. As before, all vital signs and test measures, except forWAT, OLS, and RB, were assessed after the second portion ofalcohol was consumed (second period). Consequently, subjects’BACs were measured about 2 h after starting drinking.

After the second period, subjects enjoyed snack foods and softdrinks, but no alcohol, for an additional 45 min. After this time, allvital signs and test measures, now including WAT, OLS, and RB,were assessed once more (final period). Therefore, subjects’ BACswere measured at least 1 h after drinking was finished.

Data Analysis

Data were analyzed for main effects of restedness and BAC withSPSS 17.0 (SPSS Inc., Chicago, IL). Variation of BAC within testsessions was analyzed using repeated-measures analysis of variance.Generalized estimating equations for ordinal logistic data were usedto analyze HGN and its subtests (LSP, DSNMD, ONP45), WAT,OLS clues, RB sway, RB tremor, and EN. Generalized estimatingequations for binary logistic data were used to analyze VGN andLOC. A linear mixed model was used to analyze OLS count, RBtime, systolic and diastolic BP, pulse rate, pupil size, NPC, andAFOV. Additional analyses are reported below in the respectivesections in Results.

For illustrative purposes in the figures only, data from the testsare combined for all subjects within the normal sleep and sleep-deprived conditions, and, other than the baseline results, grouped inBAC increments of 0.04 g ⁄dL. Vital sign data from overnight peri-ods are included in the respective figures for completeness but,except for pupil size, were not subjected to statistical analysis withrespect to the test session data.

Results

Figure 1 shows the average BAC for all subjects at each testperiod. The highest BAC reached by a single subject at one periodwas 0.115 g ⁄dL. There are no significant differences in BAC based

on gender, F1,27 = 2.80, p = 0.106, restedness, F1,27 = 2.16,p = 0.154, or the interaction of gender and restedness, F1,27 = 0.08,p = 0.774. As expected, BAC varies significantly with test period,F2,154 = 172.4, p < 0.0005.

Table 2 summarizes the results of the analyses of the FSTs, vitalsigns, and clinical tests described below, based on the statisticalanalyses conducted.

FSTs

Figure 2 shows the average number of HGN clues, out of amaximum of six. Consistent with previous research (56,62), sub-jects with BAC < 0.08 g ⁄ dL exhibited on average fewer than fourclues, while subjects at 0.08 g ⁄ dL and above exhibited on averagefour or more clues. Statistical analyses show that the total numberof HGN clues and the number of clues on each subtest increasewith BAC but do not vary significantly with restedness (seeTable 2). Likewise, McNemar tests of baseline data demonstratethat there is no significant difference in performance on any subtestbased on restedness: LSP, v2(1) = 1.00, p = 0.317; DSNMD,v2(1) = 1.60, p = 0.206; and ONP45, v2(1) = 0, p = 1.00.

Figure 3 shows the percentage of subjects who exhibited VGN.VGN typically is present with certain drugs (57) or at a highBAC for the individual (62) in the presence of at least four clueson the HGN test. On average, fewer than 25% of subjects withnonzero BAC under 0.08 g ⁄ dL exhibited VGN, while 40% ofsubjects with BAC at or above 0.08 g ⁄dL exhibited VGN. Statisti-cal analysis shows that the presence of VGN increases with BACbut does not vary significantly with restedness (see Table 2). Eval-uators of two of the 29 subjects (6.9%) after normal sleepobserved VGN at the baseline assessment, i.e., at BAC of0.00 g ⁄ dL. Interestingly, one of these evaluators did not observeVGN on one of the subjects after the first period of the session,and neither subject exhibited VGN at the baseline assessmentfollowing SD.

Figure 4 shows the average number of WAT clues, out of amaximum of eight. Consistent with previous research (56), subjectswith BAC < 0.08 g ⁄ dL exhibited on average fewer than two clues,while subjects at 0.08 g ⁄ dL and above exhibited on average two ormore clues. Statistical analysis shows that the number of WATclues increases with BAC but does not vary significantly with rest-edness (see Table 2).

Figure 5a shows the average number of OLS clues, out of amaximum of four. Consistent with previous research (56), subjectswith BAC < 0.08 g ⁄ dL exhibited on average fewer than two clues.Interestingly, even subjects at 0.08 g ⁄dL and above exhibited onaverage fewer than two clues, although the variances are greaterthan for subjects with BAC under 0.08 g ⁄dL. Statistical analysisshows that the number of OLS clues increases with BAC but doesnot vary significantly with restedness (see Table 2).

Figure 5b shows the average number to which subjects countedduring the OLS test. While this is not one of the validated clues ofthe OLS test, officers frequently record the results to demonstratehow suspects perform on this cognitive task, especially if suspectsmiscount or make other mistakes. Statistical analysis shows thatOLS count does not vary significantly with BAC but does decreasewith SD (see Table 2), from an overall mean (stdev) of 25.8 (3.65)when well rested to 23.8 (3.73) when sleep deprived.

Figure 6a shows the average number of RB sway and tremorclues. Statistical analysis shows that the average number of swayclues increases significantly with intoxication but does not vary sig-nificantly with restedness. Neither SD nor alcohol intoxication areexpected to cause tremors, but intoxication with other drugs (57) or

FIG. 1—Average blood alcohol concentration (BAC), in g ⁄ dL, at eachtest period by gender across sessions. Standard error bars indicated. Opensymbols: normal sleep; filled symbols: sleep deprived. Triangles: females;circles: males.



fatigue resulting from overexertion, exhaustion, or various musclediseases (64) can do so. Statistical analysis confirms that the aver-age number of tremor clues does not vary significantly with BACor restedness (see Table 2).

Figure 6b shows the average actual time elapsed when subjectsestimated the passage of 30 sec. Statistical analysis shows that RBtime increases with BAC, from a mean (stdev) of 33.53 (4.68) secfor all baseline measures to 35.52 (4.93) sec for all non-baselinemeasures. The former average is well within the normal range of25 to 35 sec (49,57). The latter average is only slightly beyond themaximum time allowed for this test, but not statistically differentfrom it, t(57) = 0.80, p = 0.214. RB time does not vary signifi-cantly with restedness (see Table 2).

Figure 7 shows the percentage of subjects who exhibited LOC.Baseline results indicate that up to about 30% of subjects could notconverge their eyes to the bridge of the nose. While this is a betterresult than expected based on the findings of a normative study(59), it supports the change in procedure to the DRE protocoldescribed above for this test. Statistical analysis shows that thepresence of LOC increases with BAC but does not vary signifi-cantly with restedness (see Table 2).

Vital Signs

Most subjects exhibited BP and pulse rate measures within orslightly below normal ranges during all test periods, including

TABLE 2—Results of statistical tests for main effects of restedness and blood alcohol concentration (BAC).

Change with Restedness? Change with BAC?

Statistic Value p Value Statistic Value p Value

Field sobriety testsHorizontal gaze nystagmus� 0.81 0.775 74.10 <0.0005*

Lack of smooth pursuit� 0.002 0.962 58.52 <0.0005*

Distinct and sustained nystagmus at maximum deviation� 0.004 0.951 38.23 <0.0005*

Onset of nystagmus prior to 45 degrees� 0.52 0.472 53.68 <0.0005*

Vertical gaze nystagmus� 0.36 0.550 38.62 <0.0005*

Walk-and-turn� 0.023 0.879 34.50 <0.0005*

One-leg standClues� 0.074 0.786 22.40 <0.0005*

Count� 11.17 (1, 77.9) 0.001* 2.20 (1, 82.5) 0.142Romberg balance

Sway� 2.00 0.157 10.61 0.001*

Tremor� 2.43 0.119 0.12 0.734Time� 1.46 (1, 85.0) 0.230 3.97 (1, 89.7) 0.049*

Lack of convergence� 0.003 0.957 7.57 0.006*

Vital signsBlood pressure

Systolic� 1.97 (1, 201.0) 0.162 0.66 (1, 203.0) 0.417Diastolic� 10.68 (1, 201.0) 0.001* 0.33 (1, 204.9) 0.564

Pulse rate� 5.36 (1, 201.0) 0.022* 25.90 (1, 203.3) <0.0005*

Pupil size� 237.1 (1, 201.0) <0.0005* 2.45 (1, 204.1) 0.119Clinical tests

Endpoint nystagmus� 1.00 0.318 14.02 <0.0005*

Nearpoint of convergence� 4.43 (1, 201.0) 0.037* 28.40 (1, 204.4) <0.0005*

Attentional field of view� 11.83 (1, 201.0) 0.001* 23.90 (1, 207.3) <0.0005*

�Wald chi-square with one degree of freedom.�F-value with degrees of freedom in parentheses.*Significant at p < 0.05.

FIG. 2—Average number of horizontal gaze nystagmus (HGN) clues atbaseline and with respect to blood alcohol concentration (BAC). Standarderror bars indicated. Open bars: normal sleep; filled bars: sleep deprived.

FIG. 3—Percentage of subjects who exhibited vertical gaze nystagmus(VGN) at baseline and with respect to blood alcohol concentration (BAC).Open bars: normal sleep; filled bars: sleep deprived.



overnight and baseline, which is no cause for concern. No abnor-mal or adverse changes were reported for any subject at any time,even for the subject with known hypertension.

Figure 8 shows the average systolic and diastolic BP. Statisticalanalyses show that systolic BP does not vary significantly with rest-edness or BAC, with an overall mean (stdev) of 121.8 (18.9)

mmHg, and that diastolic BP increases slightly but significantlywith SD, from 75.2 (10.0) to 77.8 (8.7) mmHg, and does not varysignificantly with BAC (see Table 2).

Figure 9 shows the average pulse rate. Statistical analysis showsthat pulse rate varies significantly with both restedness and BAC

FIG. 4—Average number of walk-and-turn (WAT) clues at baseline andwith respect to blood alcohol concentration (BAC). Standard error barsindicated. Open bars: normal sleep; filled bars: sleep deprived.

(a)

(b)

FIG. 5—Average number of one-leg stand (OLS) (a) clues and (b) countat baseline and with respect to blood alcohol concentration (BAC). Stan-dard error bars indicated. Open bars: normal sleep; filled bars: sleepdeprived.

(a)

(b)

FIG. 6—(a) Average number of Romberg balance (RB) sway (S) and tre-mor (T) clues at baseline and with respect to blood alcohol concentration(BAC). Standard error bars indicated. Open bars: normal sleep; filled bars:sleep deprived. (b) Average Romberg balance (RB) time estimation of thepassage of 30 sec at baseline and with respect to blood alcohol concentra-tion (BAC). Standard error bars indicated. Open bars: normal sleep; filledbars: sleep deprived.

FIG. 7—Percentage of subjects who exhibited lack of convergence (LOC)at baseline and with respect to blood alcohol concentration (BAC). Openbars: normal sleep; filled bars: sleep deprived.



(see Table 2), at baseline decreasing from a mean (stdev) of 72.4(13.2) bpm after normal sleep to 68.8 (13.2) bpm with SD, butincreasing during the test periods to overall means of 77.1 (15.2)bpm after normal sleep and 75.9 (13.9) bpm with SD.

Only one subject exhibited anisocoria of 1 mm at a single mea-sure, and only four other subjects exhibited anisocoria of 0.5 mmduring any test period, resulting in an overall prevalence of 5 of 29(17.2%) for this cohort. This is consistent with previous findingsfor both magnitude and prevalence (65). Consequently, pupil sizesare reported and analyzed only as the average over the two eyesfor each subject. Figure 10 shows average pupil size in mm. Dur-ing the overnight period, most subjects exhibited pupil sizes abovethe maximum of the normal range for ‘‘room light,’’ i.e., 5 mm. Atthe three measures taken during nighttime hours, all before 7 AM,mean (stdev) was 6.21 (0.76) mm. This is expected, given the rela-tively low lighting during the overnight periods, and is not a causefor concern. For the fourth measures during the overnight periods,all taken after 8 AM at the higher lighting level noted above, mean(stdev) was 5.66 (0.99) mm. Using a two-tailed paired t-test, this isnot statistically different than the baseline measure during the actualtest session a few hours later under slightly brighter lighting, 5.42(0.73) mm, t(56) = 1.02, p = 0.312. For the actual test sessions, sta-tistical analysis shows that overall average pupil sizes consistentlyare almost 1 mm larger with SD, from 4.47 (0.85) mm to 5.41(0.73) mm, which is statistically significant (see Table 2). Nonethe-less, average pupil sizes do not vary significantly with BAC.

Clinical Tests

Figure 11 shows the percentage of subjects who exhibited EN.Statistical analysis shows that the presence of EN does not varysignificantly with restedness but increases with BAC (see Table 2),from about 70% at baseline, which is only slightly greater thanreported elsewhere (60), to 100% at BAC of 0.08 g ⁄dL and above.

Figure 12 shows average NPC. Statistical analysis shows thatconsistent with previous research, NPC recedes significantly withboth SD (66) and BAC (67,68) (see Table 2), from <4 cm on aver-age at baseline to almost 8 cm on average at BAC of 0.08 g ⁄ dLand above.

AFOV is calculated as the total field over both eyes. Figure 13shows average AFOV. Statistical analysis shows that consistentwith previous research, AFOV is significantly reduced with bothSD (31) and BAC (30,68,69) (see Table 2), from about 94 deg onaverage at baseline to about 76 deg on average at BAC of0.08 g ⁄ dL and above.

Discussion

The presence and number of impairment clues typically assessedwith FSTs by law enforcement officers—HGN, VGN, WAT, OLS,LOC, and RB—do not increase with SD of 24 to 32 h, whereas allbut RB tremor do increase with BAC, as expected. As all subjects

FIG. 8—Average systolic (S) and diastolic (D) blood pressure (BP), inmmHg, at overnight, baseline, and with respect to blood alcohol concentra-tion (BAC). Standard error bars indicated. Striped bars: four overnight testintervals; open bars: normal sleep; filled bars: sleep deprived.

FIG. 9—Average pulse rate, in bpm, at overnight, baseline, and withrespect to blood alcohol concentration (BAC). Standard error bars indi-cated. Striped bars: four overnight test intervals; open bars: normal sleep;filled bars: sleep deprived.

FIG. 10—Average pupil size, in mm, at overnight, baseline, and withrespect to blood alcohol concentration (BAC). Standard error bars indi-cated. Striped bars: four overnight test intervals; open bars: normal sleep;filled bars: sleep deprived.

FIG. 11—Percentage of subjects who exhibited endpoint nystagmus (EN)at baseline and with respect to blood alcohol concentration (BAC). Openbars: normal sleep; filled bars: sleep deprived.



served under both states of restedness, each subject was his or herown control with regard to any potential effects of SD and alcoholintoxication. The separate assertions that SD increases the preva-lence of DSNMD (54) and reduces the angle of ONP45 (53), eachthereby potentially increasing the number of HGN clues observed,are not substantiated by the current study. The finding that SDexacerbates positional alcohol nystagmus (52) was not evaluated,because officers do not assess this response on impaired drivers(48,49).

A previous report on a single subject (51) and a recent study(42) suggest that SD causes decrements in smooth pursuit eyemovements. It is known that 10–20% of a normal population mayhave problems with smooth pursuits (42,70), depending on the testprotocol. However, it is uncertain to what extent such problemswould contribute to potential clues exhibited during the LSP sub-test. Scientific and clinical testing of patients typically is conductedin an environment and with procedures that are different than thoseencountered during a traffic stop or drug evaluation. Recordings ofeye movements made with specialized instrumentation can identifyminute changes in eye position and speed, which likely would notbe recognized by mere observation, such as during a clinicalscreening conducted by a doctor or the LSP subtest conducted byan officer. To wit, Bahill et al. (51) employed nonpredictable targetmotion, which is not consistent with the stimulus movement during

the LSP subtest (48). In addition, Fransson et al. (42) report adecrease in smooth pursuit gain of about 4% only after 36 h ofwakefulness and a decrease in smooth pursuit accuracy of about16% after 24 h of wakefulness, with an improvement in accuracyafter 36 h. Neither group of authors drew any conclusions aboutthe HGN test. Interestingly, in the current study, fewer subjects atbaseline exhibited LSP when sleep deprived (only three of 29) thanafter normal sleep (six of 29), which is counter to the implicationof the prior research.

The current study also demonstrates that SD does not have a sig-nificant effect on the other physiological and psychophysical tasksassessed with the FSTs. The small but significant reduction in OLScount can be attributed to the fact that counting is a cognitive func-tion. Nonetheless, OLS count is not a clue that is considereddirectly for the evidence of impairment caused by intoxication.

BP and pulse rate are known to increase with varying levels ofalcohol consumption (71). The low to moderate levels of alcoholintoxication incorporated within this study had no significant effecton BP but raised pulse rate slightly. The small but significantchanges with SD in diastolic BP, pulse rate, and pupil size couldbe attributed to changes in stress hormone levels (11,72), but not toany caffeine in the beverages most subjects consumed in the morn-ing before the test session (73). Nonetheless, such small changes invital signs would not raise an officer’s suspicion that a suspectcould be under the influence of an impairing drug in addition to orother than alcohol (57). We were unable to assess pupil sizes indim lighting (i.e., near-total-darkness [49]) and pupil reaction tolight, both of which are evaluated by clinicians and DREs alike;perhaps future research can address the effects of SD and intoxica-tion on these physiological responses.

Changes in the clinical assessments of EN, NPC, and AFOVwith either or both SD and intoxication could assist the clinician todistinguish these conditions from other organic or environmentalfactors.

For multiple logistical reasons, we could not hold test sessionsduring late evening or early morning hours, which would havemore directly matched the closing times of most establishments thatserve alcohol and when many traffic stops occur. However, thetimes of the test sessions of this study were similar to those used inprior research (18–21,25,26,32–37,39). In general, those studiesreached conclusions regarding SD similar or related to those foundin studies conducted in the hours around midnight (22–24,27,38).Consequently, even though the precise effects of the subjects’ circa-dian rhythms on the test measures cannot be determined in thisstudy, and especially because everyone’s circadian rhythm does notfollow the same time course, we do not believe that testing in thehours around midnight would have resulted in any different find-ings in this study. Future research could investigate this hypothesis.

Conclusion

While SD can affect cognitive ability and certain physiologicalresponses, the results of this study suggest that there is no evidencethat it affects eye movements or motor skills assessed with FSTs ina manner that would lead a law enforcement officer to concludethat the suspect is intoxicated, unless intoxication also is present.

Acknowledgments

We are grateful to Mr. Bruce Hochstein, Hillsboro (OR)Liquor Store, and Oregon Liquor Control Commission, for sup-plying the alcoholic beverages. We thank Washington County(OR) Sheriff’s Office for providing the facility for the alcohol

FIG. 12—Average nearpoint of convergence (NPC), in cm, at baselineand with respect to blood alcohol concentration (BAC). Standard error barsindicated. Open bars: normal sleep; filled bars: sleep deprived.

FIG. 13—Average attentional field of view (AFOV), in deg, at baselineand with respect to blood alcohol concentration (BAC). Standard error barsindicated. Open bars: normal sleep; filled bars: sleep deprived.



workshops; Breath Testing Section, Oregon State Police, forproviding the breath analysis instruments and calibration equip-ment; and especially, the law enforcement officers who gener-ously volunteered their time to participate in this study: Trp.Jeromy Hasenkamp, Ofc. Robert Hayes, Sgt. Mike Herb, Sgt.Mike Iwai, Sgt. Jeff Niiya, Sgt. Robert Voepel, Sgt. Steve Web-ster. Finally, we thank Dr. John R. Hayes for assistance withthe statistical analyses. The opinions expressed in this report aresolely those of the authors and do not necessarily reflect thoseof their employers or the individuals, agencies, or institutionsacknowledged.

References

1. Pack AI, Pack AM, Rodgman E, Cucchiara A, Dinges DF, Schwab CW.Characteristics of crashes attributed to the driver having fallen asleep.Accid Anal Prev 1995;27:769–75.

2. Lindsay GA, Hanks WA, Hurley RD, Dane S. Descriptive epidemiologyof dozing and driving in a college student population. J Am Coll Health1999;47:157–62.

3. Crum MC, Morrow PC, Olsgard P, Roke PJ. Truck driving environmentsand their influence on driver fatigue and crash rates; Paper No. 01-2294.Transport Res Record: J Transport Res Board 2001;1779:125–33.

4. Corfitsen MT. Tiredness! A natural explanation to the Grand Rapid‘‘DIP.’’ Accid Anal Prev 2003;35:401–6.

5. Lam LT. Factors associated with young drivers’ car crash injury: com-parisons among learner, provisional, and full licensees. Accid Anal Prev2003;35:913–20.

6. Smith R, Cook LJ, Olson LM, Reading JC, Dean JM. Trends of behav-ioral risk factors in motor vehicle crashes in Utah, 1992–1997. AccidAnal Prev 2004;36:249–55.

7. Rajaratnam SMW. Legal issues in accidents caused by sleepiness. JHum Ergol 2001;30:107–11.

8. Jones CB, Dorrian J, Rajaratnam SMW. Fatigue and the criminal law.Ind Health 2005;43:63–70.

9. Drobnich D. A National Sleep Foundation’s conference summary: thenational summit to prevent drowsy driving and a new call to action. IndHealth 2005;43:197–200.

10. Waddon A, Roberts I, Baker C. Long-distance driving: fatigue and risktaking. Police J 1990;63:224–38.

11. Howard A, Joint M. Fatigue and stress in driving. Empl Couns Today1994;6:3–7.

12. Williamson AM, Feyer A-M, Friswell R. The impact of work practiceson fatigue in long distance truck drivers. Accid Anal Prev 1996;28:709–19.

13. Feyer A-M, Williamson A, Friswell R. Balancing work and rest to com-bat driver fatigue: an investigation of two-up driving in Australia. AccidAnal Prev 1997;29:541–53.

14. Feyer A-M. Fatigue: time to recognise and deal with an old problem.BMJ 2001;322:808–9.

15. Gaba DM, Howard SK. Patient safety: fatigue among clinicians and thesafety of patients. N Engl J Med 2002;347(16):1249–55.

16. Scott LD, Hwang W-T, Rogers AE, Nysse T, Dean GE, Dinges DF. Therelationship between nurse work schedules, sleep duration, and drowsydriving. Sleep 2007;30:1801–7.

17. Tanida K. Reducing the effects of driving fatigue with the adoption of alane following assistance system. JSAE Rev 2000;21:258–60.

18. Nilsson L-G, B�ckman L, Karlson T. Priming and cued recall in elderly,alcohol intoxicated and sleep deprived subjects: a case of functionallysimilar memory deficits. Psychol Med 1989;19:423–33.

19. Dawson D, Reid K. Fatigue, alcohol and performance impairment. Nat-ure 1997;388:235.

20. Pilcher JJ, Walters AS. How sleep deprivation affects psychologicalvariables related to college students’ cognitive performance. J Am CollHealth 1997;46:121–6.

21. Lamond N, Dawson D. Quantifying the performance impairment associ-ated with fatigue. J Sleep Res 1999;8:255–62.

22. Williamson AM, Feyer A-M. Moderate sleep deprivation producesimpairments in cognitive and motor performance equivalent to legallyprescribed levels of alcohol intoxication. Occup Environ Med2000;57:649–55.

23. Mills KC, Spruill SE, Kanne RW, Parkman KM, Zhang Y. The influ-ence of stimulants, sedatives, and fatigue on tunnel vision: risk factorsfor driving and piloting. Hum Factors 2001;43(2):310–27.

24. Williamson AM, Feyer A-M, Mattick RP, Friswell R, Finlay-Brown S.Developing measures of fatigue using an alcohol comparison to validatethe effects of fatigue on performance. Accid Anal Prev 2001;33:313–26.

25. Elmenhorst D, Elmenhorst E-M, Luks N, Maass H, Mueller E-W,Vejvoda M, et al. Performance impairment during four days partial sleepdeprivation compared with the acute effects of alcohol and hypoxia.Sleep Med 2009;10:189–97.

26. Arnedt JT, Wilde GJS, Munt PW, MacLean AW. How do prolongedwakefulness and alcohol compare in the decrements they produce on asimulated driving task? Accid Anal Prev 2001;33:337–44.

27. Roach GD, Dorrian J, Fletcher A, Dawson D. Comparing the effects offatigue and alcohol consumption on locomotive engineers’ performancein a rail simulator. J Hum Ergol 2001;30:125–30.

28. Philip P, Taillard J, Sagaspe P, Valtat C, Sanchez-Ortuno M, Moore N,et al. Age, performance and sleep deprivation. Sleep Res 2004;13:105–10.

29. Ball K, Owsley C, Sloane ME, Roenker DL, Bruni JR. Visual attentionproblems as a predictor of vehicle crashes in older drivers. InvestOphthalmol Vis Sci 1993;34:3110–23.

30. Moskowitz H, Sharma S. Effects of alcohol on peripheral vision as afunction of attention. Hum Factors 1974;16:174–80.

31. Rog� J, P�bayle T, El Hannachi S, Muzet A. Effect of sleep deprivationand driving duration on the useful visual field in younger and oldersubjects during simulator driving. Vision Res 2003;43:1465–72.

32. Huntley MS, Centybear TM. Alcohol, sleep deprivation, and drivingspeed effects upon control use during driving. Hum Factors 1974;16:19–28.

33. Hack MA, Choi SJ, Vijayapalan P, Davies RJO, Stradling JR. Compari-son of the effects of sleep deprivation, alcohol and obstructive sleepapnoea (OSA) on simulated steering performance. Respir Med2001;95:594–601.

34. Schulte T, M�ller-Oehring EM, Strasburger H, Warzel H, Sabel BA.Acute effects of alcohol on divided and covert attention in men.Psychopharmocol 2001;154:61–9.

35. Krull KR, Smith LT, Sinha R, Parsons OA. Simple reaction time event-related potentials: effects of alcohol and sleep deprivation. Alcohol ClinExp Res 1993;4:771–7.

36. Fairclough SH, Graham R. Impairment of driving performance causedby sleep deprivation or alcohol: a comparative study. Hum Factors1999;41(1):118–28.

37. Horne JA, Reyner LA, Barrett PR. Driving impairment due to sleepinessis exacerbated by low alcohol intake. Occup Environ Med 2003;60:689–92.

38. Banks S, Catcheside P, Lack L, Grunstein RR, McEvoy RD. Lowlevels of alcohol impair driving simulator performance and reduceperception of crash risk in partially sleep deprived subjects. Sleep2004;27:1063–7.

39. Vakulin A, Baulk SD, Catcheside PG, Anderson R, van den Heuvel CJ,Banks S, et al. Effects of moderate sleep deprivation and low-dose alco-hol on driving simulator performance and perception in young men.Sleep 2007;30:1327–33.

40. Howard ME, Jackson ML, Kennedy GA, Swann P, Barnes M, PierceRJ. The interactive effects of extended wakefulness and low-dose alco-hol on simulated driving and vigilance. Sleep 2007;30:1334–40.

41. Rowland LM, Thomas ML, Thorne DR, Sing HC, Krichmar JL, DavisHQ, et al. Oculomotor responses during partial and total sleep depriva-tion. Aviat Space Environ Med 2005;76:C104–13.

42. Fransson PA, Patel M, Magnusson M, Berg S, Almbladh P, Gomez S.Effects of 24-hour and 36-hour sleep deprivation on smooth pursuit andsaccadic eye movements. J Vestib Res 2008;18:209–22.

43. Morad Y, Lemberg H, Yofe N, Dagan Y. Pupillography as an objectiveindicator of fatigue. Curr Eye Res 2000;21:535–42.

44. Russo M, Thomas M, Thorne D, Sing H, Redmond D, Rowland L, et al.Oculomotor impairment during chronic partial sleep deprivation. ClinNeurophysiol 2003;114:723–36.

45. Franzen PL, Siegle GJ, Buysse DJ. Relationships between affect, vigi-lance, and sleepiness following sleep deprivation. J Sleep Res 2008;17:34–41.

46. Franzen PL, Buysse DJ, Dahl RE, Thompson W, Siegle GJ. Sleep depri-vation alters pupillary reactivity to emotional stimuli in healthy youngadults. Biol Psychol 2009;80:300–5.

47. Summala H, Hakkanen H, Mikkola T, Sinkkonen J. Task effects on fati-gue symptoms in overnight driving. Ergonomics 1999;42:798–806.

48. National Highway Traffic Safety Administration. DWI detection andstandardized field sobriety testing student manual, 2006. PublicationDOT HS 178 R8 ⁄ 06.



49. National Highway Traffic Safety Administration. Drug evaluation andclassification training, drug recognition expert school student manual,2006. Publication DOT HS 172 R12 ⁄ 06.

50. Suzuki J. Signs related to abnormal eye movements. In: Uemura T, Su-zuki J, Hozawa J, Highstein SM, editors. Neuro-otological examination.Baltimore, MD: University Park, 1977;39.

51. Bahill AT, Iandolo MJ, Troost BT. Smooth pursuit eye movements inresponse to unpredictable target waveforms. Vision Res 1980;20:923–31.

52. Hill RJ, Collins WE, Schroeder DJ. Influence of alcohol on positionalnystagmus over 32-hour periods. Ann Otol 1973;82:103–10.

53. Tharp VK. Gaze nystagmus as a roadside sobriety test. Abstr Rev Alco-hol Driv 1981;2:5–8.

54. Booker JL. End-position nystagmus as an indicator of ethanol intoxica-tion. Sci Justice 2001;41:113–6.

55. Schapero M, Cline D, Hofstetter HW. Dictionary of visual science, 2ndedn. Philadelphia, PA: Chilton, 1968;496.

56. Stuster J. Validation of the standardized field sobriety test battery at0.08% blood alcohol concentration. Hum Factors 2006;48:608–14.

57. Kosnoski EM, Yolton RL, Citek K, Hayes CE, Evans RB. The drugevaluation classification program: using ocular and other signs to detectdrug intoxication. J Am Optom Assoc 1998;69:211–27.

58. National Highway Traffic Safety Administration. Drug evaluation andclassification training, drug recognition expert school student manual,2004. Publication DOT HS 172 R08 ⁄ 04.

59. Scheiman M, Gallaway M, Frantz KA, Peters RJ, Hatch S, Cuff M,et al. Nearpoint of convergence: test procedure, target selection, and nor-mative data. Optom Vis Sci 2003;80:214–25.

60. Abel LA, Parker L, Daroff RB, Dell’Osso LF. End-point nystagmus.Invest Ophthalmol Vis Sci 1978;17:539–44.

61. Shallo-Hoffmann J, Schwarze H, Simonsz HJ, Muhlendyck H. A reex-amination of end-point and rebound nystagmus in normals. Invest Oph-thalmol Vis Sci 1990;31:388–92.

62. Citek K, Ball B, Rutledge DA. Nystagmus testing in intoxicated individ-uals. Optometry 2003;74:695–710.

63. Jones AW. Disposition and fate of ethanol in the body. In: Garriott JC,editor. Medical-legal aspects of alcohol, 4th edn. Tucson, AZ: Lawyers& Judges, 2003;65.

64. Engel AG. Diseases of muscle (myopathies) and neuromuscular junction.In: Wyngaarden JB, Smith LH, editors. Cecil textbook of medicine, 18thedn. Philadelphia, PA: WB Saunders, 1988;2271.

65. Lam BL, Thompson HS, Corbett JJ. The prevalence of simple anisoco-ria. Am J Ophthalmol 1987;104:69–73.

66. von Noorden GF. Binocular vision and ocular motility: theory and man-agement of strabismus, 4th edn. St. Louis, MO: CV Mosby, 1990;429.

67. Wilson G, Mitchell R. The effect of alcohol on the visual and ocularmotor systems. Aust J Ophthalmol 1983;11:315–9.

68. Hill JC, Toffolon G. Effect of alcohol on sensory and sensorimotorvisual functions. J Stud Alcohol 1990;51:108–13.

69. Gustafson R. Visual attentional span as a function of a small dose ofalcohol. Percept Mot Skills 1986;63:367–70.

70. Trillenberg P, Lencer R, Heide W. Eye movements and psychiatric dis-ease. Curr Opin Neurol 2004;17:43–7.

71. Sepp� K, Sillanaukee P. Binge drinking and ambulatory blood pressure.Hypertension 1999;33:79–82.

72. Tsopanakis C, Tsopanakis A. Stress hormonal factors, fatigue, and anti-oxidant responses to prolonged speed driving. Pharmacol BiochemBehav 1998;60:747–51.

73. Waring WS, Goudsmit J, Marwick J, Webb DJ, Maxwell SR. Acute caf-feine intake influences central more than peripheral blood pressure inyoung adults. Am J Hypertens 2003;16(11 Pt 1):919–24.

Additional information—reprints not available from author:Karl Citek, O.D., Ph.D.Professor of OptometryPacific University College of Optometry2043 College WayForest Grove, OR 97116E-mail: [email protected]



mdja understanding the horizontal gaze nystagmus

Documents