(A)

(B)

(D)

(C)

(E)

faces objects

normal orientation

inverted(blocks 3 &

4 only)

target-foil examples

ROI Exp. group Control t pM (sd) M (sd)

Eyes

Mouth

Nose

Face bndry

Background

33.0% (17.5)

10.4% (5.1)

30.1% (16.3)

17.0% (11.6)

5.4% (3.8)

44.9% (23.4)

4.8% (6.3)

35.9% (19.7)

8.9% (12.7)

4.8% (4.1)

1.164

-1.984

.639

-1.331

-.318

.132a

.034a

.533b

.103a

.377a

a) one-sided test, b) two-sided test

ROI Exp. group Control t pM (sd) M (sd)

Eyes

Mouth

Nose

Face bndry

Background

29.1% (11.5)

25.3% (22.2)

20.3% (9.8)

17.0% (4.4)

17.2% (3.0)

35.3% (12.7)

14.8% (10.6)

26.9% (10.9)

16.3% (8.6)

11.5% (5.8)

1.029

-1.146

1.282

-.214

-2.512

.161a

.142a

.221b

.417a

.016a

a) one-sided test, b) two-sided test

ADI-R social & gaze at eyes

0%

10%

20%

30%

40%

50%

60%

0 5 10 15 20 25 30

ADI-R B score

fixat

ion

time

with

in R

OI

included: r = -.747 excluded: r = -.726

ADI-R social & gaze at eyes

0%

10%

20%

30%

40%

50%

60%

70%

80%

0 5 10 15 20 25 30

ADI-R B score

initi

al fi

xatio

ns w

ithin

RO

I

included: r = -.832 excluded: r = -.832

Distribution of initial fixations across ROIs

0%

10%

20%

30%

40%

50%

60%

Boundary Eyes Mouth Nose Background

Region of Interest

# of

initi

al fi

xatio

ns

control

experimental

*

Distribution of gaze across ROIs

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

Boundary Eyes Mouth Nose Background

Region of Interest

fixat

ion

time

with

in R

OI

control

experimental

*

Gaze at eyes and ADI-RC.Y. Trepagnier1*, M.M. Sebrechts1, R. Ramloll2, M. Coleman1, A. Finkelmeyer1, L. Barker2

M. Jones Paxton1, K. Gleeson1, W. Stewart1

(1) Department of Psychology, The Catholic University of America, Washington, DC, 20064 (2) National Rehabilitation Hospital, Washington, DC 20010

Partial support for this work was provided by the National Institute on Disability and Rehabilitation Research (NIDRR), US Department of Education, grant # H133E980025.

ABSTRACTIn an earlier study using a head-worn stereoscopic display and eye tracker, high-

functioning Autism Spectrum Disorder (ASD) participants looked less than controls at the interior of the face (Trepagnier et al., 2002). We report here preliminary results for a similar task using a desk-top tracker and monitor display.

Of the 9 experimental group participants whose data have been analyzed to date (mean age 17,5, range 12,1 to 22,8, all with a clinical diagnosis of Aspergers Disorder, 2 female), correlations for gaze duration and for number of initial fixations on eyes with the ADI-R social subscore were -.747 and -.832, respectively.

Seven of the 9 met Autism Diagnostic Interview (ADI-R) social subscore criterion for ASD. For these 7, correlations were -.726 and -.832 (all correlations significant at p < .05). A trend in the same direction for total ADI-R score failed to reach significance.

In contrast, the overall pattern of gaze was similar for the two groups. Experimental group participants differed from controls only in that they produced significantly longer fixations at the background and more initial fixations on the mouth.

Apparatus & Stimuli:(A) - 24" CRT TV, 30" away from participant (B) - ISCAN desktop eye tracker, 60Hz sampling frequency; 1 deg accuracy; dark pupil method

(C) - Intergraph workstation (stimulus presentation and data recording) & (D) - Pentium I PC (control of eye tracker)

(E) - high-backed chair (head stabilized by air-evacuated polystyrene-bead cushion)

Stimuli:o pictures of faces and objectso faces subtend same visual angle as actual

faces at that distanceo half inverted in blocks 3 & 4

Procedure & Measures:Recognition task:o 4 blocks of 24 trialso 8 acqusition trials followed by 16 recog-

nition trials: 8 repeats, and 8 foils (see examples on the right)

o each stimulus was presented for 4 secondso task: indicate by "yes" or "no" whether each

test picture previously seeno experimenter signals beginning of verbal

response via keyboard; then codes response as yes, no, or other

o fixation target centered horizontally near bottom of screen prior to each picture

o order of stimuli within block fixed, pseudo-randomized

o order of blocks counter-balanced; blocks 1 & 2 always before blocks 3 & 4

Additional measures:o Picture Peabody Vocabulary Test (PPVT) -

both groupso Autism Diagnostic Interview (ADI-R) -

experimental group only

Calibration:o 5-point calibrationo checked using two concentric circles around

targets, diameter 0.5o & 1o

o Calibration repeated if seemed possible to improve or if measure beyond outer circle

o Recalibrated between each pair of blocks

BACKGROUNDReduced face-gaze relative to controls has been documented in visual records

of infants later diagnosed autistic (Kubicek, 1979; Sparling, 1991; Osterling & Dawson, 1994), and empirical studies have suggested atypical distribution of face gaze in older children (Langdell, 1978; Hobson, et al., 1988). Recently video eye tracking has been used to measure localization of autistic face gaze, with findings of reduced attention to the eyes (Klin et al., 2002) and to central face (Trepagnier et al., 2002). Klin and colleagues also report that preferential attention to the mouth predicts better social functioning. Other investigators have found little or no difference between autistic and non-autistic participants (van der Geest et al., 2002). We present preliminary results from a continuation of the study reported in

Trepagnier et al. (2002) using two-dimensional (as opposed to stereoscopic) display.

Hypothesis:

We hypothesized that individuals with ASD would differ from controls matched for chronological and verbal mental age in that they would show less well-organized face gaze, in particular less gaze at the eyes and at the central face, and more gaze at the mouth and the (featureless) background.We also wished to examine the relationship between face gaze distribution

and development of the social impairments.

METHODS

RESULTS & DISCUSSIONGaze data analysis:1. raw sensory data are transformed into screen coordinates by applying appropriate linear

transformations; off-screen coordinates are treated as missing data �=> this yields points of regard (PORs)

2. PORs are combined into fixations of min duration 100 ms, with maximum distance of 0.5 degrees visual angle from the fixation center �=> this yields fixations an associated duration, and coordinates of the fixation center

3. Fixations whose centers fall within predefined regions of interest (ROIs) are identified

Group comparisons:(Test phase, upright faces. Participants not meeting social criterion excluded)

A) % total fixation time within ROIs (during entire trial, 4s):

B) initial fixation after trial onset: �(expressed as % total number of recorded initial fixations)

Correlation with ADI-R social subscore:A) % total fixation time within 'Eyes' ROI:

Discussion:Over the 4 seconds of exposure, Experimental group participants' larger amount of time

spent on the background is consistent with reduced organization. Directing initial gaze at the mouth is consistent with earlier findings of attention to this area. There is the possibility, however, that the location of the fixation square, centered at bottom of screen, artifactually added to this result.

More surprising is the finding that although the Experimental group's reduced gaze at the eyes did not reach significance, the inverse correlation within the group between gaze at eyes and ADI-R social score is large and significant. The ADI-R addresses early behavior, with special emphasis on the fifth year of life. Individuals' behavior during childhood, then, appears to be related to the extent to which they attempt to recognize faces by looking at the eyes. While these data are provided by adolescents and young adults, it appears to be the case that face-gaze begins to be established in infancy (LeGrand et al., 2001). It is speculated that the gaze behavior displayed in this study is a glimpse back in time to the underpinnings of the social impairments that were evident by the developmental periods addressed in the ADI-R.

At the same time, the pattern of gaze distribution of the Experimental group is not grossly different, consistent with the fact that the functioning level of these individuals is overall very high. It is not, then, surprising that some studies have failed to find difference, particularly if gaze data are acquired during a task that does not motivate the individual to attempt to remember, recognize or otherwise attend to the face.

B) # initial fixations within 'Eyes' ROI:

Participants:Experimental group:n=9, mean age=17.4 years (12, 1 - 22, 8), 7 male, 2 female; clinical diagnosis of Aspergers Disorder

Control group:n=9, mean age=18.4 years (12, 8 - 23, 4), 5 male, 4 female

CONCLUSION:In an earlier paper (Trepagnier, 1996) it was proposed that failure to establish face-gaze in

early life, because of an anxiety and/or mood-disorder liability that might reduce the infant's ability to cope with intense social stimulation, could of itself account for much of the social and communicative impairment of autism. The preliminary data reported here are modestly consistent with that hypothesis. The hypothesis will be tested directly if an intervention, now in development, to stimulate development of face attention and face interpretation proves workable.

References:Hobson, R. P., et al. (1988). What's in a face? The case of autism. British J Psychol 79, 441-453.Klin, A., et al. (2002). Visual fixation patterns during viewing of naturalistic social situations as predictors of social competence in individuals with autism. Arch Gen Psychiatry 59, 809-16.Kubicek, L. F. (1979). Organization in two mother-infant interactions involving a normal infant and his fraternal twin brother who was later diagnosed autistic. Biennial Mtg Soc Res Child Dev, San Francisco.Langdell, T. (1978). Recognition of faces: an approach to the study of autism. J Child Psychol Psychiatry 19, 255-268.LeGrand R., et al. (2001). Early visual experience and face processing. Devel Science 4, 233-51.Osterling, J., & Dawson, G. (1994). Early recognition of children with autism: a study of first birthday home videotapes. J Autism Dev Dis 24, 247-58.Sparling, J. W. (1991). Brief report: a prospective case report of infantile autism from pregnancy to four years. J Autism Dev Dis 21, 229-36.Trepagnier, C. (1996). A Possible Origin for the Social and Communicative Deficits of Autism. Focus on Autism and Other Dev Dis 11, 170-82.Trepagnier, C., et al. (2002) Atypical face gaze in autism. Cyberpsychol Behav 5, 213-7.van der Geest et al. (2002) Gaze behavior of children with pervasive developmental disorder toward human faces: a fixation time study. J Child Psychol Psychiat 43, 669-78