paradoxical functional facilitation in brain-behaviour ... · paradoxical functional facilitation...

16
Brain (1996), 119, 1775-1790 REVIEW ARTICLE Paradoxical functional facilitation in brain-behaviour research A critical review Narinder Kapur Wessex Neurological Centre, Southampton; Department of Correspondence to: Dr N. Kapur, Wessex Neurological Psychology, University of Southampton Centre, Southampton General Hospital, Southampton SO]6 6YD, UK Summary The aim in this review article is to document research findings that have shown paradoxical effects of nervous system changes, whereby direct or indirect neural damage may result in facilitation of behavioural functions. Such findings have often been ignored or undervalued in the brain- behaviour research literature. A further aim is to consider possible mechanisms and theoretical insights related to this facilitation. Analyses of relevant studies show that two major types of paradoxical functional facilitation (PFF) effects may be distinguished, (i) Situations where damage to intact brain tissue brings to normal or near normal a previously sub- normal or abnormal level offunctioning. I refer to improved levels of functioning in such contexts as restorative PFF effects. One of the best documented examples of such PFF effects is the 'Sprague effect', whereby collicular lesions may bring about an improvement in visual functioning following an initial occipital lesion, (ii) Situations where a subject with nervous system pathology or sensory loss performs better than normal control subjects on a particular task. I refer to improved levels ofperformance in these contexts as enhancing PFF effects. Restorative and enhancing PFF effects have Keywords: paradoxical improvement; functional facilitation Abbreviation: PFF = paradoxical functional facilitation been found in a range of domains, including memory, sensory and perceptual functions, and language functioning. A potential contribution of PFF effects is that they highlight two important neural mechanisms, i.e. inhibition and compensatory plasticity. Two broad classes of theoretical insights related to PFF effects are therefore discussed: (i) inhibitory mechanisms, which form part of an interactive view of brain function where competitive opponent-processing is a significant feature; (ii) 'compensatory augmentation', which occurs as a specific manifestation of CNS plasticity. Both of these mechanisms are considered in relation to paradoxical increases in CBF and anatomical annexation effects that are seen in neurological patients and in subjects with sensory loss. Paradoxical functional facilitation paradigms represent a powerful methodological tool for confirming or refuting hypotheses in brain-behaviour research. The counter-intuitive nature of PFF findings provides a particularly persuasive set of evidence in support of neural, conceptual or computational models of brain function that specifically predict paradoxical facilitation of cognitive functioning. Introduction When we are faced with a subject who has a brain lesion, it naturally follows that emphasis is placed on the deficits which the subject displays. This is for at least two reasons: firstly, trying to alleviate/analyse these deficits in clinical contexts is usually the main purpose of the interaction © Oxford University Press 1996 between the clinician/researcher and a patient; secondly, it is usually the deficits in neurological function which are more prominent compared with other features of a subject's condition. Thus, if the subject does have some change in signs or symptoms which represent an improvement rather

Upload: buiquynh

Post on 02-May-2018

234 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

Brain (1996), 119, 1775-1790

REVIEW ARTICLE

Paradoxical functional facilitation inbrain-behaviour researchA critical review

Narinder Kapur

Wessex Neurological Centre, Southampton; Department of Correspondence to: Dr N. Kapur, Wessex NeurologicalPsychology, University of Southampton Centre, Southampton General Hospital, Southampton SO]6

6YD, UK

SummaryThe aim in this review article is to document research findingsthat have shown paradoxical effects of nervous systemchanges, whereby direct or indirect neural damage mayresult in facilitation of behavioural functions. Such findingshave often been ignored or undervalued in the brain-behaviour research literature. A further aim is to considerpossible mechanisms and theoretical insights related to thisfacilitation. Analyses of relevant studies show that two majortypes of paradoxical functional facilitation (PFF) effects maybe distinguished, (i) Situations where damage to intact braintissue brings to normal or near normal a previously sub-normal or abnormal level of functioning. I refer to improvedlevels of functioning in such contexts as restorative PFFeffects. One of the best documented examples of such PFFeffects is the 'Sprague effect', whereby collicular lesions maybring about an improvement in visual functioning followingan initial occipital lesion, (ii) Situations where a subject withnervous system pathology or sensory loss performs betterthan normal control subjects on a particular task. I refer toimproved levels of performance in these contexts as enhancingPFF effects. Restorative and enhancing PFF effects have

Keywords: paradoxical improvement; functional facilitation

Abbreviation: PFF = paradoxical functional facilitation

been found in a range of domains, including memory, sensoryand perceptual functions, and language functioning. Apotential contribution of PFF effects is that they highlighttwo important neural mechanisms, i.e. inhibition andcompensatory plasticity. Two broad classes of theoreticalinsights related to PFF effects are therefore discussed:(i) inhibitory mechanisms, which form part of an interactiveview of brain function where competitive opponent-processingis a significant feature; (ii) 'compensatory augmentation',which occurs as a specific manifestation of CNS plasticity.Both of these mechanisms are considered in relation toparadoxical increases in CBF and anatomical annexationeffects that are seen in neurological patients and in subjectswith sensory loss. Paradoxical functional facilitationparadigms represent a powerful methodological tool forconfirming or refuting hypotheses in brain-behaviourresearch. The counter-intuitive nature of PFF findingsprovides a particularly persuasive set of evidence in supportof neural, conceptual or computational models of brainfunction that specifically predict paradoxical facilitation ofcognitive functioning.

IntroductionWhen we are faced with a subject who has a brain lesion, itnaturally follows that emphasis is placed on the deficitswhich the subject displays. This is for at least two reasons:firstly, trying to alleviate/analyse these deficits in clinicalcontexts is usually the main purpose of the interaction

© Oxford University Press 1996

between the clinician/researcher and a patient; secondly, it isusually the deficits in neurological function which are moreprominent compared with other features of a subject'scondition. Thus, if the subject does have some change insigns or symptoms which represent an improvement rather

Page 2: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

1776 N. Kapur

than a deficit in function, this tends to be overlooked inclinical and research contexts.

In this article, my purpose is to review those studies wheresubjects with lesions of the CNS have, in certain contexts,shown non-trivial facilitation of psychological functioning,and to consider whether such paradoxical functionalfacilitation (PFF) may offer a useful theoretical approach tothe study of neural systems. A large number of relevantobservations are scattered about in the literature. While anumber of these are well-known and firmly establishedphenomena, other related findings have tended to beoverlooked. To my knowledge, such observations have neverbeen integrated into one comprehensive review article.

By 'paradoxical' I mean improved performance on testsof cerebral function, where this improvement is in theopposite direction to detrimental effects that have usuallyaccompanied the neural changes in question. In this review,I will be primarily concerned with changes induced byexperimental brain lesion, naturally occurring disease or bysensory deprivation. Psychological manipulations that resultin PFF effects are rare (e.g. Boiler and Marcie, 1978), andare somewhat outside the scope of this article. Certainparadoxical effects in human and animal psycho-pharmacology, such as the retrograde memory facilitationeffects of compounds that usually have anterograde amnesiceffects (Parker and Weingartner, 1984), are also outside thescope of this article.

In considering improvements in functioning after cerebralpathology, I will be focusing on two distinctive scenarios,(i) Situations where alteration to intact neural tissue bringsto normal or near-normal a level of functioning that waspreviously sub-normal or abnormal. In these instances, thereare real physical, cognitive or other.psychological gains as aresult of the' specific neural alteration/damage that has beensustained. I refer to improved levels of functioning in suchcontexts as restorative PFF effects, although the degree ofimprovement may not always reflect a restoration tocompletely normal levels of functioning, (ii) Situations wherea patient with a damaged nervous system can perform betterthan normal control subjects on a particular task. This taskmay have been especially designed and constructed to takeaccount of the patient's functional deficits. I refer to improvedlevels of performance in such contexts as enhancing PFFeffects.

I will be mainly concentrating on situations where PFFeffects are non-trivial and are not simply secondary to taskinstructions or to extensive practice of a particular pieceof behaviour. The phenomenon of so-called idiots savantsrepresents a somewhat unusual type of enhanced performancethat could be interpreted in this latter way (Treffert, 1988).It seems that much of the 'better-than-normal' behaviour insuch subjects can simply be explained by repeated practice,often to an obsessional degree, of a particular set ofbehaviours. Thus, the condition of idiots savants, which hasbeen extensively reviewed in its own right, will not beconsidered here.

Some human developmental studies have indicatedsituations where children perform better than adults. Forexample, Werker et al. (1981) have shown that infants areable to discriminate non-native speech sounds (e.g. Hindispeech sounds) better than English speaking adults,presumably because the infants' speech perception systemshave not yet been 'primed' with the biasing effect of theexperience of hearing native, English speech sounds. Thereis also evidence from animal studies of enhanced learningby infant animals compared with adult animals (Spear et al.,1994). At the neurobiological level there is evidence in non-human primates for elimination and refinement of brainprojections during the progression from infancy to adulthood(Rodman, 1994). However, since this article is primarilyconcerned with acute changes related to the occurrence ofspecific types of neuropathology in the adult brain, I will notbe giving such developmental studies further consideration.

Restorative PFF effectsIn this section I will concentrate on PFF effects that entailimprovements in cognitive functions after cerebral pathology.Restorative PFF effects have already been well documentedin clinical conditions such as epilepsy (Purves, 1991) andmovement disorder (Marsden and Obeso, 1994) [see Ianseket al. (1995) and Marsden and Obeso (1995) for furtherdiscussion of paradoxical facilitation effects in the area ofmovement disorder]. I will also not be covering the realm ofbeneficial personality changes after brain pathology whichhas, by its nature, been subject to less experimental rigourthan the cognitive areas examined below. Changes intemperament and personality have often been based onclinical observations rather than objective personalitymeasures. In addition, improvements in temperament/moodstate after brain pathology may occur at a cost, either interms of other personality changes that may be maladaptive(e.g. the person may now become rather placid and apathetic),or in reduced cognitive functioning. Beneficial changes inpersonality after brain surgery have also been well coveredelsewhere (e.g. Hitchcock et al., 1979; Bartlett, 1987; Waziri,1990; Diering and Bell, 1991).

PerceptionIn a seminal paper, Sprague (1966) observed a functionalfacilitation effect in cats whereby a left-sided hemianopiabrought about by removal of the right occipito-temporalneocortex could be replaced by evidence of normal responsesto stimuli in the left visual field as the result of subsequentdestruction of the left superior colliculus. A similar effectwas obtained when the commissure that connected the twocolliculi was sectioned, though the return of vision in theleft visual field was less rapid than where the colliculusitself had been destroyed. Sprague argued that the initialhemianopia was due to depression of function of the colliculusipsilateral to the cortical lesion, and that this depression was

Page 3: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

Paradoxical functional facilitation 1777

maintained by influx of inhibition from the other colliculus,the one contralateral to the cortical lesion. The removal ofthat colliculus, or sectioning of the collicular commissure,was considered to have abolished this inhibition. Therefore,return of function in the ipsilateral colliculus led to somedegree of recovery from the hemianopia. The degree ofrecovery appeared to vary between cats, and appeared torange from complete to partial recovery, but the importantpoint to note is that significant recovery did take place. Itseems that the initial hemianopia may have reflected bothloss of cortically mediated vision and also loss of vision inthe midbrain that was mediated by damage to cortical-midbrain fibre connections. Sprague concluded that, 'Thehemianopia that follows unilateral removal of the cortex thatmediates visual behaviour cannot be explained simply inclassical terms of interruption of the visual radiations thatserve cortical function. Explanation of this deficit requires abroader point of view, namely, that visual attention andperception are mediated at both forebrain and midbrain levels,which interact in their control of visually guided behaviour.Hemianopia caused by cortical lesion is due to an imbalanceof these neural centres that subserve vision, resulting in analteration of function at the midbrain level. Imbalance canbe redressed and vision restored to the previously hemianopicfield by subsequent lesion in the superior colliculus' (Sprague,1966, p. 1547). Sprague's findings, which have come to beknown as the 'Sprague effect', were subsequently confirmedby Sherman (1974), by Kirvel et al. (1974) and by Hardyand Stein (1988). Sherman (1977) showed the effect occurredregardless of the sequence of cortical and collicular lesions.Goodale (1973) was able to produce a similar type of twin-lesion effect in respect of light-evoked potential recordedfrom the superior colliculus in the rat. Wallace et al. (1989)showed that the recovery of visual function in the 'Spragueeffect' is not mediated by fibres arising from the contralateralsuperior colliculus but rather by fibres in the caudal half ofthe commissure that connected the two superior colliculi.Wallace et al. (1990) found that the 'Sprague effect' couldalso be produced by destruction of neurons that gave rise toaxons in the caudal collicular commissure. These neuronswere primarily located in the rostro-lateral substantia nigrapars reticulata.

A further PFF effect has recently been reported by Sprague(1991). He first showed that splitting the optic chiasmand introducing a cortical, suprasylvian lesion in the righthemisphere interfered with the transfer of form perceptionfrom the left to the right eye. He then demonstrated thatsectioning of the superior collicular commissure, or of thecolliculus that was contralateral to the side of the suprasylvianlesion, could restore inter-hemispheric transfer of formperception that had been disrupted by the presence of thechiasm lesion and the suprasylvian lesion.

More recently, Ptito et al. (1993) reported a PFF effectsimilar to that originally reported by Sprague in his 1966paper. Cats with impaired visual functioning following lesionsof the left optic tract and right visual cortex showed a

significant improvement of retention and learning of visualpattern discriminations after subsequent sectioning of thecorpus callosum. Ptito et al. (1993) argued that thehemispheric disconnection interrupted abnormal influencesexerted by the lesioned hemisphere on the intact lefthemisphere. These abnormal influences included inhibitionof visual pathways that trajected via the collicular system. Itwas therefore presumed that sectioning of the corpus callosumhelped to remove this inhibition, and that this mediated theparadoxical improvement in performance. A further PFFeffect in the area of visual function has been noted byintroducing light deprivation after frontal lesions that inducedvisual neglect (Crowne et al., 1983; Corwin and Vargo,1993). Corwin and Vargo (1993) found that, in rats withlesions of the left medial agranular frontal cortex, severeright-sided visual neglect resulted. However, if this lesion wasfollowed by 48 h of light deprivation, there was acceleratedrecovery of function on visual and tactile neglect tasks.Corwin and Vargo were unable to explain their findings,but hypothesized that light deprivation may somehow havereduced an imbalance in activation between the twohemispheres, an imbalance brought about by the originallesion, and that the environmental manipulation may haveaffected dopaminergic pathways, which had previously beenassociated with recovery from neglect (Corwin et al., 1986;Fleet etal., 1987).

Human analogues of the Sprague effect have rarely beenreported—perhaps the closest was a recent paper (Vuilleumieret al., 1996), in which the authors reported the disappearanceof left-sided unilateral neglect, brought on by a right parietalinfarct, after the occurrence of a second lesion in the area ofthe left frontal eye field. Using the rationale offered by theSprague effect, two groups of investigators have attemptedto initiate improvements in neglect in human subjects.Following a suggestion by Posner and Rafal (1987), Butterand Kirsch (1992) reasoned that, by patching the right eyeof patients with left-sided neglect, the left superior colliculuswas deprived of its major facilitatory input. Therefore, theinhibition that it can exert on the right superior colliculuswill be reduced. This should then enable the right superiorcolliculus to orient more readily towards the left side ofspace, with a consequent improvement in neglect. Butter andKirsch (1992) were able to produce improvement in a numberof neglect tasks by devices such as patching the right eye inpatients with left-sided neglect. However, Soroker et al.(1994) were only able to produce improvement in neglect inone out of six patients after introducing ipsilesional eyepatching. Although, in a subsequent study, the sameinvestigators (Serfaty et al., 1995) did find improvement in13 out of 26 patients, two patients also benefited from contra-lesional monocular occlusion, and the authors concluded thatthe Sprague effect was unlikely to be the sole factor involvedin the facilitation that had been observed in some patients.

Speech expressionThere are at least three separate reports of patients whostuttered since childhood, and in whom there was a cessation

Page 4: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

1778 N. Kapur

of stuttering after the onset of specific cerebral pathology. Inall of these patients where handedness was documented, thisyielded evidence of left/mixed handedness in the patient orin his immediate family. Jones (1966) reported three casesin which surgery for the removal of a left hemisphere lesionin adulthood resulted in the cessation of stuttering and onecase where removal of a right hemisphere lesion gave thesame benefit. The lesions in question were either an aneurysmor a cerebral tumour. On the basis of sodium amytal studies,he noted that, whereas prior to brain surgery speech appearedto be bilaterally represented, after removal of the lesion inthe damaged hemisphere, the speech seemed to be representedsolely in the intact hemisphere. This was the right hemispherein three patients and the left hemisphere in the remainingpatient. Thus, it appeared that prior to surgery bothhemispheres were involved in the imperfect regulation ofspeech. After surgery, with only one hemisphere subservingspeech, it now seemed that, as far as speech fluency wasconcerned, one hemisphere was better than two!

Miller (1985) described two patients whose stutteringceased when they developed early signs of multiple sclerosis.In both patients, there was evidence of bilateral cerebellardysfunction, and the author considered that this might,paradoxically, have been responsible for the dramaticcessation of the stuttering.

Helm-Estabrooks et al. (1986) reported the case of anambidextrous man who had begun to stutter at the age of 8years. When he was an adult, he ceased to stutter after theoccurrence of a severe head injury in which a right hemisphereclot was removed. This improvement was evident as soon ashe recovered consciousness, 10 days after the injury. Oneyear later, his speech was still much improved comparedwith what it was prior to the head injury, with only someminor problems in articulation and fluency.

More recently, it has been shown that left thalamicstimulation alleviated stuttering in four patients with evidenceof 'subcortical seizure activity' and a history of chronic painin addition to their speech disturbance (Andy and Bhatnagar,1992). Although none of these patients had evidence ofstructural lesions in the thalamus, it is unclear if the'subcortical seizure activity' represented some abnormalityin the thalamus or in related pathways.

Enhancing PFF effectsSensory and perceptual functioningBefore discussing lesion studies, I will firstly consider casesof indirect neural changes brought about by sensory loss.The possible enhancement of sensory functioning in onemodality after total sensory loss in another modality has beenthe subject of many studies over the years. For the purposesof this paper, I will restrict discussion mainly to recent studiesrelating to sensory functioning in the blind, although similarobservations have been made in other sensory deprivedsubjects. As noted by Bach-y-Rita (1994, p. 459), 'blindness

can be considered to result in 'brain damage', since the lossof a major sensory input, such as vision, markedly alterscortical activity'.

Rauschecker and Kniepert (1994) found that visuallydeprived cats, who had both eyelids sutured from birth, wereable to localize a sound in space more precisely thannormal control cats. Korte and Rauschecker (1993) foundthat auditory spatial tuning of single neurons in the anteriorectosylvian cortex was sharper in visually deprived cats thannormal control cats. This sharpening of spatial tuning wasfound across the whole frequency range. At the tactile level,Rauschecker et al. (1992) have reported hypertrophy offacial vibrissae in visually deprived cats and mice, witha concomitant enlargement of their representation in thesomatosensory cortex.

Muchnik et al. (1991) presented several auditory tasks toblind and sighted human subjects, including tasks coveringauditory localization, temporal discrimination and speechperception. They found that blind subjects performed betterthan normal control subjects in all three tasks. As Muchniket al. (1991), Rauschecker (1995) and Ronnberg (1995) pointout in their reviews of earlier work, the overall picture withregard to enhanced auditory functioning in blind subjectsremains a complex one, with some divergence betweenstudies. In the case of olfaction, there is an equivalent degreeof divergence between studies (for a review of earlierwork, see Smith et al. 1993). Two of the better controlledinvestigations have produced rather different outcomes.Murphy and Cain (1986) found that blind subjects performedbetter than sighted subjects on an odour identification task,where they had to recall the name of an odour. Using asimilar task, however, where recognition rather than recallof names was tested, Smith et al. (1993) did not findany difference between sighted and control subjects. Theyattributed the contrast with the Murphy and Cain (1986)results to the greater sensitivity of the recall task used byMurphy and Cain. In the case of surface texture, Walker andMoylan (1994) found that blind subjects showed better recallthan normal control subjects of the surface texture of objectsthat had earlier been described using the auditory modality.There was evidence that this superiority emerged graduallyin blind subjects in proportion to the percentage of theirlifetime that was spent without sight.

The interpretation of any enhancement in sensory functionin sensorily deprived subjects is of course a matter of debate.Simple training/practice effects are usually contrasted withmore spontaneous mechanisms such as changes in inhibitory-excitatory balance, neuronal reorganization, neuronal sprout-ing, etc. At present, there does not appear to be any firmevidence in support of one or other explanation, and itremains possible that a matrix of mechanisms interact toinfluence a particular form of sensory enhancement.

One of the first perceptual studies in which enhancing PFFeffects after specific cerebral pathology were reported, is onethat was essentially a problem solving experiment but whereperceptual impairments may have contributed towards the

Page 5: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

Paradoxical functional facilitation 1779

facilitation effects that were observed. Golding (1981)administered a perceptual problem solving task to patientswith right and left hemisphere lesions and to control subjects.The task required subjects initially to inspect a stimulus arraywhich included the colours yellow and green, each pairedwith diamond or circle shapes. Either half of the card couldinclude a colour or shape, and there were therefore eightcards displayed altogether. In the criterion task, two coloursand two shapes were shown, with the other halves of thecards masked. Subjects had to indicate which cards, out offour alternative stimulus pairs, were critical for falsifying thestatement: 'whenever there is a circle on one half of the card,there is a yellow colour on the other half of the card'. Morethan half of the right hemisphere lesidhed patients, but noneof the control subjects, were able to solve the task correctly(by choosing the circle and green cards), and it transpiredthat all of those right hemisphere lesioned patients whosolved the task were ones who had difficulties in namingphotographs from unusual angles. Golding speculated thatcontrol subjects failed to solve the problem because of thedominating effects of the perceptual aspects of the task atthe expense of verbal reasoning, and that this type ofperceptual processing did not interfere with the reasoningprocesses of right hemisphere lesioned patients withperceptual classification deficits. However, Evans and Dennis(1982) argued that the particular effect which Golding foundcould have occurred simply due to the patients with righthemisphere lesions showing unilateral neglect for the leftside of the array, and consequently greater attention to stimulion the right hand side, since one of the critical stimuli (thecard with one half green) was located on the right hand sideof the array.

Ladavas et al. (1990) found that patients with left-sidedneglect due to right hemisphere lesions were faster thancontrol patients in their response to one of two horizontalstimuli presented in the ipsi-lesional (right) field—namely,the right stimulus compared with the left stimulus. Theauthors argued that, while control patients distributed theirattention fairly evenly to both horizontal stimuli, neglectpatients selectively favoured the right stimulus due to theirleft-sided inattention, and that this resulted in their paradoxicalfinding. This particular outcome could therefore be explainedin terms of 'selective attention' that occurred in the neglectpatients but not in the control patients. It should, however,be noted that control patients in the Ladavas et al. (1990)study were patients with right hemisphere lesions who didnot suffer from any neglect, and it remains possible that bothsets of patients would have been impaired relative to normal,neurologically intact, control subjects.

Townsend and Courchesne (1994) found that autisticsubjects with parietal lobe volume loss showed shorterreaction time than normal control subjects in a simple targetdetection task that involved responding when one of severalsquares contained a filled circle. They interpreted thisenhanced performance as evidence to support the concept thatvisual attention can be distributed as an attention 'spotlight' or

'beam', and that a narrowed attention focus may result fromsome types of parietal lobe abnormality.

Wray et al. (1995) have reported that migraine suffererswere significantly faster than normal control subjects on two'low-level' visual perception tasks. Stimuli were presentedin one hemifield. One task involved subjects indicating inwhich half (top or bottom) a line of differing orientationappeared. The other task required subjects to indicate whichof two dots appeared first on the screen. Wray et al. (1995)hypothesized that this superiority could be explained by thegreater visual sensitivity shown by migraine sufferers, relatedto a greater excitability of visual areas of the cortex. Thesecortical areas may be the same areas that contribute to thevisual aura associated with migraine episodes.

Sensory-motor functioningIn the case of sensory-motor performance, a dramatic exampleof enhancing PFF has recently been demonstrated by Lajoieet al. (1992). Following two episodes of sensory poly-neuropathy, their patient suffered a permanent and specificloss of the large sensory myelinated fibres in both arms andlegs. Clinical investigations showed total loss of the sensesof touch, vibration, pressure and kinaesthesia in all fourlimbs. Lajoie et al. (1992) asked their patient to perform astandard mirror drawing task, where she had to trace a six-pointed star while viewing the image of the star through amirror. The task had to be performed as accurately and asquickly as possible. For the five control subjects who werealso tested, the mirror drawing task posed difficulties, sincereversed visual information coded within a visual map hadto be compared with proprioceptive information coded in anormal proprioceptive internal map. These conflicting sourcesof information resulted in difficulties in making the correctmovements, especially for oblique lines. By contrast, thedeafferented patient had no such sensory conflict, due to lossof proprioception. For this patient, the task was simply oneof visual tracking. Thus, she performed significantly betterthan normal control subjects on the mirror drawing task. Thisphenomenon could be seen as the proprioceptive equivalentof the Stroop effect, whereby an automatic encoding process(in this case, generated by proprioceptive cues in relation tomovement in space) interfered with a second process involvedin task performance (in this case, generated by visuospatialcues).

Memory functioningAnimal studiesA large number of animal lesion studies have led to reportsof enhancing PFF effects in the area of memory functioning(for a review of some of these studies, see Spear et al., 1994).Monkeys with split-brain lesions that involved sectioning ofthe corpus callosum have been shown to be able to learnconflicting visual discriminations if these are presented

Page 6: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

1780 N. Kapur

separately to each hemisphere, while this task proved to bevery difficult and frustrating for a normal control animal(Trevarthen, 1962). Some investigators have examinedreversal learning, which usually entails the animals learninga particular discrimination-or-reward-association, and thenlearning the reverse procedure in which the originally 'correct'stimulus is made 'incorrect', and vice versa. They typicallyfound impaired reversal learning in animals with lesionswithin the limbic system (e.g. Winocur and Olds, 1978, whoworked with rats). However, there are a few reports ofenhancing PFF effects which resulted in better-than-normalreversal learning by animals with limbic system lesions. Forexample, Eichenbaum et al. (1986) found superior reversallearning of odour discriminations in rats who had bilaterallesions of the fornix or combined bilateral lesions of thefornix and amygdala. In a recent study, Eichenbaum andBunsey (1995) showed that rats with hippocampal lesionswere better than control rats at distinguishing odour pairsthat were associated with a reward stimulus. These authorshypothesized that hippocampal lesioned rats relied on a'fused' representation of odour pairs, with the fusion mediatedby intact structures such as the parahippocampal gyrus,whereas control rats treated odours within a pair as discreteitems to be remembered, a strategy that resulted in slowerlearning compared with control rats.

Staubli et al. (1984) provided evidence that rats withlateral entorhinal lesions showed enhanced reversal learningcompared with control rats on an olfactory discriminationlearning task. Staubli et al. (1984) suggested that it was rapidforgetting of information from earlier trials, brought aboutby the entorhinal lesions, that contributed to the enhancedreversal learning. Zola and Mahut (1973) reported facilitationof visual reversal learning by monkeys with bilateral lesionsof the fornix. One of the reasons for this facilitation mayhave been that animals with limbic lesions may simply haveshown more rapid forgetting of information from earliertrials, with this memory loss resulting in less interferenceduring their learning of the second set of reversal trials.Zola and Mahut (1973) suggested that some form of fasterextinction of earlier responses, rather than weak learning onthe initial trials, may have contributed to the facilitationeffects which they found. It is unclear why the reversallearning data from these four studies are at variance fromthose of other published studies of the effects of limbicsystem pathology, though it is possible that cue modality anddegree of training sophistication may have been critical,especially for control animals. Shaw and Aggleton (1993)found that rats with fornix lesions performed better thancontrol rats on the acquisition trials of a delayed nonmatchingto sample task. They argued that this was due to the presenceof spatial biases in the responses of control animals, whichresulted in their failing to return to a maze arm from whichthey had come and therefore performing poorly on specifictrials where the animal was required to return to the samearm used in the previous trial. Shaw and Aggleton (1993)reasoned that fornix lesioned rats did not use such allocentric

cues to direct their performance, and that this absence of aspatial bias contributed towards their better performancecompared with control animals. Irle (1985) proposed that theslight facilitation effect found in a study of learning in catsafter extensive lesions in the limbic system may have beendue in part to a greater calmness and better attentional statein lesioned animals, compared with control animals whotended to be more active during testing.

Several other PFF effects have been reported in animallesion studies of memory functioning. Butter and Doehrman(1968) reported that monkeys with lateral striate lesionsshowed better performance on a visual discrimination learningtask than control subjects, due to an apparent reduction inthe interference they suffered from having learned similarvisual stimuli on an earlier training session. Butter andDoehrman (1968) reasoned that this paradoxical enhancementof transposition following posterior cortical lesions was dueto an impairment in detecting differences between the trainingand test stimuli. Klapdor et al. (1994) have recently shownthat bilateral lesions of the tuberomamillary nucleus, anarea in the posterior hypothalamus, resulted in enhancedperformance by rats on a one-trial step-through avoidancetask. They hypothesized that the lesions resulted in adisinhibition of reinforcement processes, possibly related tohistamingergic mechanisms.

Prado de Carvalho and Zornetzer (1981) found that ratswho were given bilateral locus coeruleus lesions prior to theacquisition of an inhibitory avoidance response showedaccelerated forgetting of the response 14 days after training.However, when the rats were then administered electro-convulsive shock at this time, this resulted in a paradoxicalimprovement in memory functioning. Galey et al. (1989)found that lesions of the septum resulted in facilitation ofspatial discrimination learning compared with controlanimals. They argued that these lesions reduced the inhibitorycontrol exerted by the septum on cholinergic neurons of thesepto-hippocampal pathway. However, as the authors pointedout, their findings did conflict with those from an earlierstudy (Simon et al., 1986).

Human studies

Alcoholic Korsakoff patients/patients with globalamnesia. Turning to human lesion studies, a number ofstudies that could be classified under the rubric of 'priming'have found enhanced performance in amnesic patients. Itshould, however, be remembered that the majority of primingstudies have found normal or below normal performancein amnesic as compared with control subjects (Graf andMasson, 1993).

Gardner et al. (1973) carried out a study in which alcoholicKorsakoff patients first learned words related to certaincategories, e.g. the name of a tree is an oak. They had tolearn six such items to a criterion of two successive correcttrials. Ninety seconds later, they were administered several

Page 7: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

Paradoxical functional facilitation 1781

retention test conditions. One of these took the form of animplicit memory test, where subjects simply had to generateitems from the earlier set of categories. Alcoholic Korsakoffpatients were more likely than alcoholic control subjects tooffer items from the target set that they originally learned.'Here, possession of a relatively intact memory leads to arecall score which is apparently lower than that of a groupcharacterized by its deranged memory' (Gardner et al.,1973, p. 173). One of several mechanisms may have beenresponsible for this finding. It is possible that the effectrelates to the population of available responses in the twogroups of subjects. For controls, this may be relatively large,but for Korsakoff patients, due to their retrograde amnesia,it may only have consisted of the pool of items which weremost prominent in their immediate past experience—in thiscase the set of words used in the earlier learning task. Asthe authors of this particular study pointed out, their resultsmay also have been due to Korsakoff and control subjectsinterpreting instructions in the verbal fluency task ratherdifferently. It is possible that control subjects thought that intheir generation of words they were to omit any responseswhich overlapped with items in the learning task that hadbeen administered a short time earlier. Korsakoff patients,on the other hand, would probably have forgotten about thememory task and just given any items which came to mind,with the recently learned items still having some trace intheir memory.

Warrington and Weiskrantz (1978) reported significantfacilitation effects in a study of word-completion performancein amnesic patients. Subjects were first asked to try togenerate words from the initial three letters of a set of words.These items were unique words, e.g. juice, ankle and aisle,such that there was only one possible answer to completethe three-letter word-stem. They were later given a memorytask, the first stage of which involved copying and readinga set of 20 whole words which the patients could not completefrom the set of 200. After 24 h, they were given an unexpectedretention test for these items, where they were presented with40 three-letter stems, 20 from the set which they copied andread, and 20 which they successfully completed in the earlierpart of the study. They were required to complete theseletter stems. Amnesic patients performed better than controlsubjects on this implicit retention task. Here again, it wouldappear that prior learning may have had a more dominatingeffect on the performance of amnesic patients compared withcontrol subjects. However, it is of note that amnesic patientswere also superior to control subjects on the very first task,i.e. generating words for the set of 200 three-letter stems,and it remains possible that this superiority also played apart in their enhanced performance on the retention test.

Jacoby and Witherspoon (1982) presented amnesic patientsand control subjects with a naming task, where the testsentence included a homophone (e.g. 'Name a musicalinstrument that employs a reed'). They then administeredan oral spelling task which included the set of 'primed'homophones. This was followed by a standard recognition

test for the set of target words. Amnesic patients were morelikely than control subjects to spell a homophone by the wayin which it was biased by the earlier naming task. Onthe recognition memory test, they displayed the expectedrecognition memory deficit. Jacoby and Witherspoon (1982)speculated that it may have been perseverative tendencies oftheir amnesic patients which led them to respond in this way,thus rendering them more sensitive to the effects of an earlierexperience. At least two qualifications in respect of theJacoby and Witherspoon study need to be stated. First, someof the test procedures were not described in detail, and itappears that the control subjects and amnesic patients mayhave differed in age and/or educational level. Secondly, twosubsequent studies, in which the same or similar paradigmswere used, have failed to replicate this facilitation effect(Cermak et al., 1986; Kapur, 1988). In the Cermak et al.(1986) study, control subjects actually showed a greaterpriming effect than Korsakoff patients.

Cermak et al. (1988) also examined word-stem completionperformance in alcoholic Korsakoff patients. Subjects werepresented with a series of word pairs, and after a 3-min filler-task were then provided with stems (first three letters) of thesecond words in the pair. These word stems were paired witheither the same words as in the presentation trial, with otherwords in the same set of items, or with new, 'distractor'words. Alcoholic Korsakoff patients were better than controlsubjects at completing word stems, i.e. the first letters ofwords used in the presentation trial. This effect was apparentacross the various conditions and did not depend on whetheror not subjects were told before the presentation trial that theirmemory would be tested. The paradoxical effect prompted theauthors of the study to comment: 'Why does the performanceof alcoholic Korsakoff patients on the word-stem completiontask actually exceed the performance of alcoholic controls?It appears that amnesics are more sensitive than controls tosemantic activation of a previously presented word' (1988,p. 322). In a further study (Cermak et al., 1992), the authorsfound that, in a task which involved primed stem completionof high frequency words, Korsakoff patients produced moreresponses from the primed list than alcoholic control subjects.The authors reasoned that alcoholic controls had moreconscious control over their performance, and that thisresulted in their sometimes choosing to complete stems withwords other than those on the study list. In this study, theamnesics seemed more bound than the control subjects torespond to the 'fluency' produced by an item's recentpresentation. Similarly, Squire and McKee (1993) showedthat amnesic subjects were more likely than control subjectsto judge previously presented non-famous names as famous,and noted that in their study 'prior events have been foundto influence amnesic patients to a greater extent than theyinfluence normal subjects' (1993, p. 428).

Musen et al. (1990) gave amnesic patients and controlsubjects a task in which they were required simply to readthe same story three successive times. The reduction in timeover these three trials was taken as a measure of implicit

Page 8: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

1782 N. Kapur

memory for the story. In one experiment, Musen et al. (1990)found that amnesic patients showed a significantly fasterlearning curve than control subjects, with greater latencygains on the third compared with the first trial. There wasalso a tendency for the same effect to occur in a secondexperiment, where repeated reading of the same story wasalso required, but in this case the differences did not reachsignificance. Musen et al. (1990) were uncertain of thereliability of their finding, but speculated that control subjectsmay have slowed their rate of reading to allow for a laterretention test that was to be given for the story, whereasamnesic patients might have forgotten that this retention testwas to occur and therefore this 'slowing down' factor wouldnot have influenced their reading speed.

Other memory disordered patients. In most of theabove articles, authors found PFF effects in studies ofanterograde amnesia. Warrington and McCarthy (1988) founda PFF effect in their study of retrograde amnesia. Theyreported that their amnesic patient was more successfulthan control subjects in producing the names of famouspersonalities in response to verbal cues (Christian name andfirst letter of surname). As Warrington and McCarthy noted,this finding may have been due, in part, to a possibleconfound resulting from the prior presentation of a similarface-naming task that was given to the amnesic patient butnot to the same control subjects. However, it is also possiblethat, when making such a verbal retrieval, control subjectsrely on a much wider repertoire of responses, whereas theamnesic patient is reliant on a more limited repertoire, onethat has been unaffected by his retrograde and anterogradeamnesia. This may mean that for a given verbal cue, controlsubjects will have to make a wider search and be moretempted to offer alternative names, whereas the search of theamnesic patient will be more constrained.

Heindel et al. (1990) found that in patients withHuntington's disease, there was, compared with controlsubjects, enhanced priming of the threshold to name fragmentsof pictures that had been shown previously. A similarobservation of 'hyperpriming' by Huntington's diseasepatients was made by Randolph (1991), this time usingword-stems rather than picture fragments. In a recent study,Jernigan (1994) has reported a positive correlation betweenthe amount of tissue loss in the caudate area, as estimatedfrom MRIs, and the degree of priming shown by Huntington'sdisease patients on a word identification task. Jernigan (1994)reasoned that impairment of perceptual and lexical processingof verbal stimuli, associated with caudate damage inHuntington's disease patients, may make such processingmore difficult and more effortful. In normal subjects, it hadbeen shown that degrading the visual features of stimuli, orusing less familiar stimuli, resulted in a larger degree ofpriming. Jernigan reasoned that in Huntington's diseasepatients, more difficult and effortful processing may thereforeresult in enhanced retention of stimuli, as shown on asubsequent word identification task.

The set of facilitation effects found in the studies reviewedabove can be explicable in part by control subjects, but notamnesic patients, being encumbered with explicit memoriesthat acted to hinder or interfere with task performance. Asimilarly induced type of facilitation effect has been reported,this time due to the reduction or elimination of interferencewhere the interference may stem from semantic rather thanepisodic memories. This results from a process, whichoperates in normal subjects, but which may be partlyoperational or non-operational in memory disordered patientsdue to a lesion that disrupts the process in question.

In the first study in which such an effect was examined,Kapur et al. (1986) reported an experimental investigationof a patient who presented with transient amnesic attackswhich subsequently developed into a form of temporal lobeepilepsy. Further details on the patient are presented in thatpaper and in a follow-up article (Kapur et al., 1989). Thepatient showed focal deficits on tests of public events, famousfaces and famous names in the presence of intact performanceon standard anterograde memory tasks. In an experimentalinvestigation of his memory functioning, he was presentedwith six names which he did not recognize. These werepaired with occupations, such that the associations to thenames were ones which might normally interfere withmemory performance (e.g. John Newcombe—singer). Thepatient's paired-associate learning performance was superiorto that of normal control subjects. In this case, it appearedthat interfering associations, which the stimuli provided forcontrol subjects, did not operate to affect the patient's memorytest performance adversely. A remarkably similar facilitationeffect has recently been reported by De Renzi and Luchelli(1993)—their patient was left with a dense retrogradeamnesia, but intact paired-associate learning ability. Whenhe was required to associate famous names with incongruentoccupations, he also performed better than control subjects.He thus showed an 'interference elimination' effect, wherebyhis performance was unaffected by the normal interferencethat would impinge on the behaviour of control subjects,due to the absence of prior associations to the famousnames used in the paired-associate learning task. In somerespects, these sets of findings represent a human analogueof the facilitation in reversal learning referred to earlier inthe review of animal studies, and to 'interference reduction'mechanisms proposed by Hinrichs et al. (1984) to explain theretrograde facilitation effects of diazepam on human memory.

Semantic processingKapur (1980) found facilitation effects in a study of semanticcategorization by a 'deep dyslexic' patient. Such patientstypically have difficulty in reading abstract and functionwords, as compared with concrete words. In addition, theycannot read 'nonsense words' and they make semanticparalexic errors when reading words aloud. In this study, thepatient was presented with a written target word, e.g. ostrich;he would then hear 14 stimulus words and was asked to

Page 9: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

Paradoxical functional facilitation 1783

indicate after each word was spoken whether it remindedhim of the target word in any way. Seven words bore arelationship to the target word and seven were unrelated toit. Control subjects made a large number of false negativeresponses to a same-category member of stimulus words e.g.in the case of ostrich, the word 'peacock'. Thus, they tendedto say that a word was not associated with the target word,whereas the paralexic patient made a relatively small numberof such errors. There was also a pathological tendency bythe paralexic patient to accept neutral words as stimuluswords to the target word such that he might accept the word'sun' as an associate word. It did not appear that this wasdirectly related to the patient's better-than-normal scores inrespect to false negative errors, as this was selective toone category of associates rather than reflecting an overallresponse bias.

An analogous semantic priming effect for indirectassociates has been reported for schizophrenic patients, andshown to be stronger than that found in control subjects.Spitzer et al. (1993) assessed schizophrenic patients andcontrol subjects using a reaction time paradigm where theprime could be directly related to the target word (e.g. hen-egg), or indirectly related to the target word (e.g. lemon-sweet). Subjects had to indicate whether the second wordwas a real word (in some conditions, the second wordwas a nonsense word). Compared with control subjects,schizophrenic patients showed enhanced facilitation, relativeto a neutral condition, for the indirect priming situation, andit was most apparent in thought-disordered schizophrenicpatients. The effect was particularly evident where the targetimmediately followed the prime. Spitzer et al. (1993)interpreted their findings as reflecting hightened activation/reduced inhibition in the semantic memory network ofschizophrenic patients. Their finding paralleled an earlierobservation of enhanced semantic priming in schizophrenicpatients by Kwapil et al. (1990).

Paradoxical functional facilitation: theoreticalmechanisms and methodological gainsThe evidence reviewed above has shown that subjects withcerebral dysfunction may, in certain contexts, show improvedlevels of functioning compared with their pre-lesioned stateor compared with normal control subjects. These observationshave been documented in a variety of disease aetiologies,with a number of animal species, over a number of lesionsites and across several experimental paradigms.

What theoretical insights derive from PFF effects? Whatnew methodological gains arise as a result of PFF findings?I will now attempt to address these issues.

Inhibitory mechanismsI propose that PFF effects provide novel insights into aframework of neural functioning where inhibition, and the

balance between inhibitory and excitatory neuronal popu-lations, play a critical role.

Restorative PFF effects point to the importance of'opponent' processing models of brain function, whereby, innormal subjects, inhibitory and excitatory mechanismsinteract in a complex harmony to bring about a specific setof behaviours. The role of inhibitory processes may becritical in mediating specific restorative PFF effects.Reversal of excessive inhibition may be important, as in thecase of facilitation of motor functioning following pallidallesions (Marsden and Obeso, 1994), and alleviation of avisual field defect after lesioning of the superior colliculus.A related mechanism is interference elimination—perhapsmost dramatically seen after sectioning of the corpuscallosum to relieve epilepsy (Purves, 1991). In somecircumstances, it may not be possible to distinguish betweenthe presence of simple inhibitory/excitatory mechanisms andthose that relate to interference/reduction of interference.From the point of view of PFF effects, the behaviouraloutcome will often be the same, regardless of which type ofmechanism may be operating.

Recent PET studies have shown that neural activityassociated with some types of cognitive activity may involveincreases in blood flow in task-related structures and paralleldecreases in blood flow in neighbouring, alternative sensorymodalities. (NB Increases in blood flow may arise fromincreases in firing either of inhibitory neurons or of excitatoryneurons.) Haxby et al. (1994) found that performance on afaces matching task or a visual location matching task wasaccompanied by increases in blood flow in visual regionssuch as the occipital-temporal or the parieto-occipital areas,and at the same time there were decreases in CBF in auditoryand somatosensory areas. Similarly, Kapur et al. (1995) notedblood flow reductions in the left superior temporal gyrus andin both inferior parietal areas during faces memory tasks.Inhibitory mechanisms underlying functional facilitation maybe related to reports of paradoxical increases in blood flow,in structures distal but connected to the lesion site, that havebeen noted in PET studies of patients with focal cerebrallesions (e.g. Weiller et al., 1992). This increase in CBF hasbeen hypothesized to be due to functional disinhibition ofstructures connected by inter- or intra-hemispheric pathwaysto the critical lesion site. In the case of functional facilitationphenomena, it is possible, for example, that the improvementsnoted in some memory studies may be due to selectivedisinhibition in particular frontal lobe structures connectedto critical limbic-diencephalic lesion sites.

Paradoxical functional facilitation data support PET data insuggesting a conceptual framework for cognitive performancethat emphasizes complex inhibitory and excitatory inter-actions between a number of diverse neural circuits, ratherthan the operation of discrete neural systems in isolation. Ifone regards the brain as a dynamic system, where specificstates of cerebral functioning derive from complex excitatoryand inhibitory interactions between neuronal populations,then it is quite possible that an abnormal state may reflect

Page 10: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

1784 N. Kapur

an imbalance between these sets of influences (cf. Kinsbourne,1970). The introduction of lesions to a dysfunctional brainmay lead to changes that partly or wholly rectify thisimbalance by creating new sets of excitatory and inhibitoryinteractions. This scenario has been succinctly summarizedby Wall (1987, p. 241): 'Lesions of the peripheral or centralnervous system may immediately unmask the existenceof previously unsuspected neural circuitry by way of thedifferential effect of the lesion on excitatory and inhibitorymechanisms. It follows also that the disappearance of afunction is not necessarily caused by the anatomicalinterruption of some input-output pathway necessary for thatfunction. Loss of function, as well as appearance of novelfunction, may result from shifts of excitation and inhibitionwhich mask or unmask the ability of existing anatomicalpathways to transmit nerve impulses'.

Allowing for an obvious degree of caution whenextrapolating to human contexts, and for appropriate ethicalconsiderations, we should nevertheless take more seriouslythe possibility that for some perceptual/cognitive disorderstreatment may be accelerated by the introduction of a lesionin a particular area of the brain, and that in some cases certainpathologies may offer protection against other pathologies.Rather than thinking, for example, in terms of absent ordepleted neural circuits, it may be more fruitful to think interms of mechanisms such as excessive interference whichmay be due to abnormal excitation or inhibition within anetwork of neural circuits, or due to lack of synchronybetween inhibitory and excitatory systems. The question maythen be of how we can introduce a lesion to diminishthe amount of interference and ensure a more harmoniousinterplay of neural activation. For example, apart from theeye-patching studies discussed earlier (Butter and Kirsch,1992; Soroker et al., 1994), there appears to have been noattempt with nonhuman primates to consider whether the'Sprague effect' may generalize from species such as the cat,i.e. whether improvement in cortically induced loss of visualfunction may be brought about by a second, discretesubcortical lesion.

Models of memory that have inhibitory systems amongsttheir key features are now emerging (Read et al., 1994).Within the range of inhibitory phenomena, there are particulareffects, such as latent inhibition and negative priming, whichpredict better-than-normal performance in neurologicalpatients with inhibitory deficits. Both latent inhibition andnegative priming refer to the deleterious effects of priorexposure to a stimulus, in a situation where different responsedemands are made of the stimulus than were made in theinitial priming phase. The phenomenon of negative priminghas been examined in more detail in the human population,and examples from this effect are therefore more readilyavailable (McDowd et al., 1995; Neill et al.. 1995). Theexperimental paradigm used in negative priming studiesusually involves two phases. In phase one, a target and adistractor stimulus are presented to a subject, who is askedto respond to the target stimulus but to ignore the distractor

stimulus. On a following trial, the previous distractor stimulusis now presented alone, as a target stimulus. In normalsubjects, a response to this stimulus is now slowed downbecause of persisting 'inhibitory effects' from its earlierpresentation. However, where the subject suffers from adefect in inhibition, then the negative effects of the priorexposure of the stimulus are minimized. Supporting datahave been gathered, mainly from older versus youngersubjects, with older subjects showing evidence of inhibitorydeficits that result in reduced levels of negative priming. AsMcDowd et al. (1995, p. 387) noted: 'It is interesting to notethat like latent inhibition, one of the useful aspects of thenegative-priming paradigm is that an inhibitory deficit resultsin better than normal performance. An individual withdecreased inhibitory function will actually respond faster inthe negative-priming condition than will an individual withnormal inhibitory function.' It is possible that if we can reacha level of understanding of neural mechanisms for areas suchas memory, similar to that which has already been reachedfor motor functioning (Marsden and Obeso, 1994), then thereare more realistic prospects for more novel and more soundlybased interventions that may improve cognition.

Compensatory plasticityA second major insight that is offered by PFF effects relatesto various manifestations of CNS plasticity, and in particular'compensatory augmentation'. In the case of enhancedsensory functioning in visually deprived subjects, which wasreviewed earlier, several studies have used functional imagingtechniques, either event-related potentials (Alho et al., 1993)or magnetic encephalographic responses (Kujala et al., 1995),to show that in blind subjects the 'visual' cortex is used toperform auditory detection tasks. Uhl et al. (1991) and Rosieret al. (1993) also found similar 'anatomical annexation' whenthey gave a tactile task to blind and sighted subjects;they reported higher event-related DC potential shift in theoccipital cortex area of blind subjects than in sighted subjects.Pascual-Leone and Torres (1993) reported an expansion ofthe sensorimotor cortical representation of the Braille readingfinger in blind subjects. A number of non-human studieshave shown that enucleation or thalamic lesions will resultin sprouting of new fibres that project to visual cortical areas(e.g. Cusick and Lund, 1982). While the precise nature ofthe changes reported in the above papers remains to bedetermined, it would appear that expansion or reorganizationof the cortical representation of sensory functions may takeplace after the onset of a sensory loss such as blindness, andthat, in some cases, this may play a part in the enhancednon-visual sensory functioning shown by blind subjects.

At present, it is impossible to say whether this 'compensatoryaugmentation' simply derives from the fact that sensory-deprived subjects take part in intensive practice in particularsensory activities, or whether it is due to more subtle CNSmechanisms (Ramachandran, 1993; Rauschecker, 1995).These mechanisms may include: (i) unmasking of 'silent

Page 11: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

Paradoxical functional facilitation 1785

synapses' or pre-existent 'latent' anatomical connections,perhaps as a result of disinhibition; (ii) reduced competition/interference between cortical-sensory areas; (iii) synapticsupersensitivity; or (iv) new connections being formed in theform of 'anatomical sprouting'. This sprouting may reflectcompensatory sprouting from the uninjured branches of aneuron's axonal tree or the collateral sprouting of newconnections from the axonal terminals of nearby neurons(Devor, 1994). Prolonged, intensive experience using aparticular sensory modality may result in specific structuralchanges in the brain. Alternatively, the occurrence of a directbrain lesion or deafferentiation may result in more rapidstructural reorganization at the neuronal level (cf. Chino et ai,1992; Gilbert and Wiesel, 1992). A third possibility is that bothfactors may occur to varying degrees in a given individual. Itis important to keep in mind that any compensatory structuralchanges, either at the level of specific cortical representationsor in terms of more general changes in inhibitory-excitatorybalance, may be mediated by other 'superordinate' changes.These may, for example, occur at the subcortical level or theymay be reflected in neurochemical changes. Either of thesecould then modulate changes at a hierarchically more specificlevel of neural functioning.

The extent of neuroanatomical plasticity that is possiblein altered brain states has been illustrated by observationssuch as visual responsiveness of somatosensory neurons inhamsters in whom retinal cells were made to project to themain thalamic somatosensory nucleus at birth (Metin andFrost, 1989), and the responsiveness to auditory orsomatosensory cues of the ectosylvian visual area of catswho have been deprived of vision for several years(Rauschecker and Korte, 1993). Ramachandran (1994) hassuggested that sensory remapping at the cortical levelmay explain the distribution of reference fields on the facefor phantom limb sensations that are experienced by somelimb amputees (Yang et ai, 1994). Ramachandran (1994, p.316) further predicted that in such amputees the increased'magnification factor' for the face might result in enhancedtactile perception on the faces of these subjects.

Paradoxical increases in blood flow in neurologicallydamaged patients has been observed in several PET studies,although the precise significance of these remains uncertain.Grady et ai (1993) reported that during a face-matchingtask, patients with Alzheimer-type dementia showed greaterblood flow activation than control subjects in the occipitalpole and in the frontal cortex, in the area of the frontal eyefields. One possibility is that for patients, the face-matchingwas more difficult than for control subjects, and thus requiredmore effort or greater eye-movement activity, which in turnled to increased blood flow. Weiller et ai (1992) assessedpatients with left-sided subcortical, striatocapsular infarcts.During a resting state, there were predictable reductions inblood flow in areas such as the basal ganglia and thalamus, butthese were accompanied by increased flow in the contralateralposterior cingulate and premotor cortices, and in the caudatenucleus ipsilateral to the recovered hand. Weiller et ai (1992)

also gave their patients a simple motor task to perform withtheir recovered right hand. Compared with control subjects,these patients showed increased blood flow in both insulae,in the inferior parietal, prefrontal and anterior cingulatecortices, in the ipsilateral premotor cortex and basal ganglia,and in the contralateral cerebellum. In a study of patientsduring their migraine attack, Weiller et ai (1995) foundincreased blood flow in the cingulate region, in auditory andvisual association cortices and in the brainstem. Engelienet ai (1995) observed that a patient who partially recoveredfrom auditory agnosia showed blood flow increases in rightfronto-parietal structures, that were not found in normalcontrol subjects, during a sound categorization task. In arecent PET study, Becker et ai (1996) found that, in a verbalmemory task, patients with Alzheimer-type dementia showedtwo areas of activation not found in control subjects, one inthe dorsolateral prefrontal cortex of both hemispheres andthe other in the left supramarginal and angular gyri.

Brain-behaviour methodologyFrom the methodological point of view, there are both generaland specific implications of PFF effects.

At the general level, PFF effects represent some of themost powerful forms of evidence to confirm or refutehypotheses in cognitive neuropsychology. This is becauseone has avoided the whole argument inherent in the 'deficit'approach, namely that a demonstrated impairment in a groupof neurological patients may have been simply due to a'brain-damage' effect which depressed test scores. In thisrespect, PFF effects can be seen to be an enhanced form of'double-dissociation' effects. The traditional use of double-dissociation methodology has been shown to have a numberof limitations, including the susceptibility of its outcomes tobe influenced by ceiling or floor effects (Weisktrantz, 1968;Jones, 1983; Miller, 1993) and the possibility that single-process inferences may be compatible with some double-dissociation findings (Dunn and Kirsner, 1988). Miller (1993)has also pointed to selection and statistical artefacts thatmay accompany some forms of double dissociation outcomesin single-case studies. Paradoxical functional facilitationeffects largely avoid such problems of interpretation. Thepresence of PFF effects may represent a particularly powerfulset of data with which to help decide a specific conceptualissue. One of the values of such demonstrations lies in thepower that their counter-intuitive nature gives to the designof a particular experimental paradigm, and the ability of sucha paradigm to address a particular hypothesis or conceptualissue with challenging data. Thus, enhancing PFF effectsmay serve to encourage a 'paradigm shift' (Kuhn, 1970) inneural systems research, a shift to experimental designs thatmay not only yield more convincing data but also suggestnovel ways of conceptualizing brain-behaviour relationships.

At a more specific level, there are a number of novelparadigms in cognitive psychology that may be associatedwith the occurrence of PFF effects. Such paradigms have

Page 12: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

1786 N. Kapur

been in use, to some extent, in a few studies, but hithertotheir potential has probably been underestimated. The firstparadigm is one which is similar to the one used by Kapuret al. (1986), and by other researchers in more recent years(e.g. De Renzi and Lucchelli, 1993; Young et al., 1986).This involves paired-associate learning where some of theitems are non-veridical (i.e. 'false') and some are veridical('true'). Thus, one may be examining face-name learningwhere famous personalities form the set of stimulus items,and where 'true'/'untrue' names are paired with the faces.With this type of paradigm, one can investigate topics suchas implicit learning for the items in question. In some cases,where there are strong interference effects in normal subjects,it may be possible to demonstrate significant functionalfacilitation effects.

The second paradigm is also a form of interferenceparadigm, and is partly similar to the 'stroop' task usedin a number of studies (e.g. Golden, 1976). It differs fromthe first paradigm in being less dependent on singularmemories for the source of interference, and being moredependent on generic knowledge. Essentially, the prototypeparadigm would be concerned with the operation of twocognitive processes. A task would be constructed, suchthat both processes may be sampled, and such that eitherone may have deleterious effects on performance, dependingon the instructions given to the subject. An example ofsuch a task is the naming of pictures with and withoutthe interfering effects of superimposed printed names(Lupker, 1979), though a variety of other paradigms couldbe envisaged. Here again, implicit perception/memory couldbe examined and, in some neurological patients, such asthose with visual agnosia or acquired dyslexia, significantfacilitation effects could be uncovered. Both of the aboveparadigms may be particularly applicable to patients withsemantic memory loss, such that an absent piece ofknowledge in a neurological patient will result in reductionor elimination of interference effects, and consequentlysuperior performance compared with control subjects. Athird paradigm, which remains to be exploited to producefunctional facilitation effects, is the 'oppositional' paradigmdescribed by Jacoby et al. (1992). Using this paradigm,Jacoby and other researchers have used processes whichmay be opposing each other in a particular patient for aspecific cognitive task, such as explicit recollection forprevious material on the one hand and unconscious primingeffects for the same material on the other hand. It ispossible that with the use of such a paradigm, amnesicpatients may, in some settings, show facilitation effectscompared with normal control subjects, analogous to theinterference effects that have already been demonstratedto result from amnesic patients showing greater effects ofprior experience on performance (Squire and McKee,1993). A fourth paradigm is the negative priming paradigmreferred to earlier, one which has recently been reviewedin detail elsewhere (May et al., 1995).

ConclusionsIn summary, PFF effects contribute towards five potentialadvances in our knowledge of brain function. First, theydemonstrate that inhibitory mechanisms are important forunderstanding brain-behaviour relationships. Secondly,they highlight the role of compensatory augmentation asa significant mechanism in CNS plasticity. Thirdly, PFFparadigms represent a powerful methodological tool forconfirming or refuting hypotheses in brain-behaviourresearch. Fourthly, PFF effects provide a particularlypersuasive set of evidence for models of brain function—thoseneural, conceptual or computational models that specific-ally predict PFF effects have added value compared withother models. Fifthly, PFF effects help to focus research onother paradoxical mechanisms by which lesions result inbeneficial changes to brain and to behaviour, such as instanceswhere increased lesion size may be associated with betterrecovery of function (Irle, 1987).

In one of his last papers, the late Norman Geschwind(Geschwind, 1985) referred to neglected work on paradoxicalrecovery of paralysis of the diaphragm, i.e. loss of functioncaused by sectioning of the upper cervical cord could berestored by cutting the opposite phrenic nerve. Geschwind(1985, p. 1) commented: 'This important, though neglected,experiment illustrates several important principles. First, theremust be many other cases in which the capacity for recoveryis latent, and revealed only by some further manipulation,but experimenters have only rarely been zealous in theirsearch for the right manoeuvre. Secondly, plasticity is notconfined to more advanced structures, but occurs even atlower levels. This experiment also provides evidence againsta semi-philosophical argument advanced as an objection totherapy (e.g. thalamotomy for parkinsonism) by means ofthe placement of additional lesions, that is, that it makes nosense to try to treat a damaged nervous system by damagingit further.' Geschwind's comments themselves have beensomewhat neglected. I hope that researchers and cliniciansalike will be stimulated to pursue further paradoxes, similarto those reviewed in this article.

Acknowledgements

I wish to thank John Aggleton, Chris Colbourn, Chris Frith,Stevan Harnad, Andy Young and the anonymous referees fortheir helpful comments.

References

Alho K, Kujala T, Paavilainen P, Summala H, Naatanen R. Auditor)'processing in visual brain areas of the early blind: evidencefrom event-related potentials. Electroencephalogr Clin Neurophysiol1993; 86: 418-27.

Andy OJ, Bhatnagar SC. Stuttering acquired from subcortical

Page 13: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

Paradoxical functional facilitation 1787

pathologies and its alleviation from thalamic perturbation. BrainLang 1992; 42: 385-401.

Bach-y-Rita P. Sensory substitution, volume transmission andrehabilitation: emerging concepts. In: Illis LS, editor. Neurologicalrehabilitation. 2nd ed. Oxford: Blackwell Scientific 1994: 457-68.

Bartlett JR. Psychosurgery. What proof is there of its value? In:Warlow C, Garfield J, editors. More dilemmas in the managementof the neurological patient. Edinburgh: Churchill Livingstone, 1987:139^4.

Becker JT, Mintun MA, Aleva K, Wiseman MB, Nichols T, DeKoskyST. Compensatory reallocation of brain resources supporting verbalepisodic memory in Alzheimer's disease. Neurology 1996; 46:692-700.

Boiler F, Marcie P. Possible role of abnormal auditory feedback inconduction aphasia. Neuropsychologia 1978; 16: 521—4.

Butter CM, Doehrman SR. Size discrimination and transposition inmonkeys with striate and temporal lesions. Cortex 1968; 4: 35-46.

Butter CM, Kirsch N. Combined and separate effects of eye patchingand visual stimulation on unilateral neglect following stroke. ArchPhys Med Rehabil 1992; 73: 1133-9.

Cermak LS, O'Connor M, Talbot N. Biasing of alcoholic Korsakoffpatients' semantic memory. J Clin Exp Neuropsychol 1986; 8:543-55.

Cermak LS, Bleich RP, Blackford SP. Deficits in the implicitretention of new associations by alcoholic Korsakoff patients. BrainCogn 1988; 7: 312-23.

Cermak LS, Verfaellie M, Sweeney M, Jacoby LL. Fluency versusconscious recollection in the word completion performance ofamnesic patients. Brain Cogn 1992; 20: 367-77.

Chino YM, Kaas JH, Smith EL, Langston AL, Cheng H. Rapidreorganization of cortical maps in adult cats following restricteddeafferentiation in retina. Vision Res 1992; 32: 789-96.

Corwin JV, Vargo JM. Light deprivation produces acceleratedbehavioral recovery of function from neglect produced by unilateralmedial agranular prefrontal cortex lesions in rats. Behav Brain Res1993; 56: 187-96.

Corwin JV, Kanter S, Watson RT, Heilman KM, Valenstein E,Hashimoto A. Apomorphine has a therapeutic effect on neglectproduced by unilateral dorsomedial prefrontal cortex lesions in rats.Exp Neurol 1986; 94: 683-9.

Crowne DP, Richardson CM, Ward G. Brief deprivation of visionafter unilateral lesions of the frontal eye field prevents contralateralinattention. Science 1983; 220: 527-30.

Cusick CG, Lund RD. Modification of visual callosal projectionsin rats. J Comp Neurol 1982; 212: 385-98.

De Renzi E, Lucchelli F. Dense retrograde amnesia, intact learningcapability and abnormal forgetting rate: a consolidation deficit?Cortex 1993; 29: 449-66.

Devor M. Plasticity in the neonatal and adult nervous system.In: Illis LS, editor. Neurological rehabilitation. 2nd ed. Oxford:Blackwell Scientific, 1994: 59-81.

Diering SL, Bell WO. Functional neurosurgery for psychiatric

disorders: a historical perspective. Stereotact Funct Neurosurg 1991;57: 175-94.

Dunn JC, Kirsner K. Discovering functionally independent mentalprocesses: the principle of reversed association. Psychol Rev 1988;95: 91-101.

Eichenbaum H, Bunsey B. On the binding of associations inmemory: clues from studies on the role of the hippocampal regionin paired-associate learning. Current Directions in PsychologicalScience 1995; 4: 19-23.

Eichenbaum H, Fagan A, Cohen NJ. Normal olfactory discriminationlearning set and facilitation of reversal learning after medial-temporal damage in rats: implications for an account of preservedlearning abilities in amnesia. J Neurosci 1986; 6: 1876-84.

Engelien A, Silbersweig D, Stern E, Huber W, Doring W, Frith C,et al. The functional anatomy of recovery from auditory agnosia.Brain 1995; 118: 1395-409.

Evans JS, Dennis I. Brain lesions and reasoning: a note on Golding.Cortex 1982; 18: 317-18.

Fleet WS, Valenstein E, Watson RT, Heilman KM. Dopamineagonist therapy for neglect in humans. Neurology 1987; 37: 1765-70.

Galey D, Toumane A, Durkin T, Jaffard R. In vivo modulation ofsepto-hippocampal cholinergic activity in mice: relationships withspatial reference and working memory performance. Behav BrainRes 1989; 32: 163-72.

Gardner H, Boiler F, Moreines J, Butters N. Retrieving informationfrom Korsakoff patients: effects of categorical cues and referenceto the task. Cortex 1973; 9: 165-75.

Geschwind N. Mechanisms of change after brain lesions. [Review].Ann N Y Acad Sci 1985; 457: 1-11. .

Gilbert CD, Wiesel TN. Receptive field dynamics in adult primaryvisual cortex. Nature 1992; 356: 150-2.

Golden CJ. Identification of brain disorders by the Stroop Colorand Word Test. J Clin Psychol 1976; 32: 654-8.

Golding E. The effect of unilateral brain lesion on reasoning. Cortex1981; 17: 31-40.

Goodale MA. Cortico-tectal and intertectal modulation of visualresponses in the rat's superior colliculus. Exp Brain Res 1973; 17:75-86.

Grady CL, Haxby JV, Horwitz B, Gillette J, Salerno JA, Gonzalez-Aviles A, et al. Activation of cerebral blood flow during avisuoperceptual task in patients with Alzheimer-type dementia.Neurobiol Aging 1993; 14: 35-44.

Graf P, Masson MEJ, editors. Implicit memory: new directionsin cognition, development, and neuropsychology. Hillsdale (NJ):Lawrence Erlbaum, 1993.

Hardy SC, Stein BE. Small lateral suprasylvian cortex lesionsproduce visual neglect and decreased visual activity in the superiorcolliculus. J Comp Neurol 1988; 273: 527^2.

Haxby JV, Horwitz B, Ungerleider LG, Maisog JM, Pietrini P,Grady CL. The functional organization of human extrastriate cortex:a PET-rCBF study of selective attention to faces and locations. JNeurosci 1994; 14: 6336-53.

Page 14: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

1788 N. Kapur

Heindel WC, Salmon DP, Butters N. Pictorial priming and cuedrecall in Alzheimer's and Huntington's disease. Brain Cogn 1990;13: 282-95.

Helm-Estabrooks N, Yeo R, Geschwind N, Freedman M, WeinsteinC. Stuttering: disappearance and reappearance with acquired brainlesions. Neurology 1986; 36: 1109-12.

Hinrichs JV, Ghoneim MM, Mewaldt SP. Diazepam and memory:retrograde facilitation produced by interference reduction.Psychopharmacology (Berl) 1984; 84: 158-62.

Hitchcock ER, Ballantine HT, Meyerson BA, editors. Modernconcepts in psychiatric surgery. Amsterdam: Elsevier, 1979.

lansek R, Bradshaw J, Phillips J, Morris ME, Cunnington R. Reviewarticle: the functions of the basal ganglia and the paradox ofstereotaxic surgery in Parkinson's disease [letter]. Brain 1995; 118:1613-15.

Irle E. Combined lesions of septum, amygdala, hippocampus,anterior thalamus, mamillary bodies and cingulate and subicularcortex fail to impair the acquisition of complex learning tasks. ExpBrain Res 1985; 58: 346-61.

Irle E. Lesion size and recovery of function: some new perspectives.[Review]. Brain Res Rev 1987; 434: 307-20.

Jacoby LL, Witherspoon D. Remembering without awareness. Can JPsychol 1982; 36: 300-24.

Jacoby LL, Lindsay DS, Toth JP. Unconscious influences revealed[comment]. Am Psychol 1992; 47: 802-9. Comment on: Am Psychol1992; 47: 766-79.

Jernigan T. Magnetic resonance imaging and memory disorders. In:Cermak LS, editor. Neuropsychological explorations of memoryand cognition. New York: Plenum Press, 1994: 147-57.

Jones RK. Observations on stammering after localized cerebralinjury. J Neurol Neurosurg Psychiatry 1966; 29: 192-5.

Jones GV. On double dissociation of function. Neuropsychologia1983; 21: 397-400.

Kapur N. Recognition of word associates in semantic paralexia. BrJ Psychol 1980; 71: 401-5.

Kapur N. Selective sparing of memory functioning in a patient withamnesia following herpes encephalitis. Brain Cogn 1988; 8: 77-90.

Kapur N, Heath P, Meudell P, Kennedy P. Amnesia can facilitatememory performance: evidence from a patient with dissociatedretrograde amnesia. Neuropsychologia 1986; 24: 215-21.

Kapur N, Young A, Bateman D, Kennedy P. Focal retrogradeamnesia: a long term clinical and neuropsychological follow-up.Cortex 1989; 25: 387^02.

Kapur N, Friston KJ, Young A, Frith CD, Frackowiak RSJ.Activation of human hippocampal formation during memory' forfaces: a PET study. Cortex 1995; 31: 99-108.

Kinsbourne M. A model for the mechanism of unilateral neglect ofspace. Trans Am Neurol Assoc 1970; 95: 143-6.

Kirvel RD, Greenfield RA, Meyer DR. Multimodal sensory neglectin rats with radical unilateral posterior isocortical and superiorcollicular ablations. J Comp Physiol Psychol 1974; 87: 156-62.

Klapdor K, Hasenohrl RU, Huston JP. Facilitation of learning inadult and aged rats following bilateral lesions of thetuberomammillary nucleus region. Behav Brain Res 1994: 61:113-16.

Korte M, Rauschecker JP. Auditory spatial tuning of cortical neuronsis sharpened in cats with early blindness. J Neurophysiol 1993; 70:1717-21.

Kuhn TS. The structure of scientific revolutions. 2nd ed. Chicago:University of Chicago Press, 1970.

Kujala T, Huotilainen M, Sinkkonen J, Ahonen AI, Alho K,Hamalainen MS, et al. Visual cortex activation in blind humansduring sound discrimination. Neurosci Lett 1995; 183: 143-6.

Kwapil TR, Hegley DC, Chapman LJ, Chapman JP. Facilitation ofword recognition by semantic priming in schizophrenia. J AbnormPsychol 1990; 99: 215-21.

Ladavas E, Petronio A, Umilta C. The deployment of visual attentionin the intact field of hemineglect patients. Cortex 1990; 26: 307-17.

Lajoie Y, Paillard J, Teasdale N, Bard C, Fleury M, Forget R, et al.Mirror drawing in a deafferented patient and normal subjects:visuoproprioceptive conflict. Neurology 1992; 42: 1104—6.

Lupker SJ. The semantic nature of response competition in thepicture-word interference task. Mem Cognit 1979; 7: 485-95.

Marsden CDM, Obeso JA. The functions of the basal ganglia andthe paradox of stereotaxic surgery in Parkinson's disease [seecomments]. [Review]. Brain 1994; 117: 877-97. Comment in: Brain1995; 118: 822.

Marsden CD, Obeso JA. Review article: the functions of the basalganglia and the paradox of stereotaxic surgery in Parkinson's disease[letter]. Brain 1995; 118: 1615-17.

May CP, Kane MJ, Hasher L. Determinants of negative priming.[Review]. Psychol Bull 1995; 118: 35-54.

McDowd JM, Oseas-Kreger DM, Filion DL. Inhibitory processesin cognition and aging. In: Dempster FN, Brainerd CJ, editors.Interference and inhibition in cognition. San Diego: Academic Press,1995: 363^00.

Metin C, Frost DO. Visual responses of neurons in somatosensorycortex of hamsters with experimentally induced retinal projectionsto somatosensory thalamus. Proc Natl Acad Sci USA 1989; 86:357-61.

Miller AE. Cessation of stuttering with progressive multiplesclerosis. Neurology 1985; 35: 1341-3.

Miller E. Dissociating single cases in neuropsychology. Br J ClinPsychol 1993; 32: 155-67.

Muchnik C, Efrati M, Nemeth E, Malin M, Hildesheimer M. Centralauditory skills in blind and sighted subjects. Scand Audiol 1991;20: 19-23.

Murphy C, Cain WS. Odor identification: the blind are better.Physiol Behav 1986: 37: 177-80.

Musen G, Shimamura AP, Squire LR. Intact text-specific readingskill in amnesia. J Exp Psychol Learn Mem Cogn 1990; 16: 1068-76.

Neill WT, Valdes LA, Terry KM. Selective attention and theinhibitory control of cognition. In: Dempster FN, Brainerd CJ,

Page 15: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

Paradoxical functional facilitation 1789

editors. Interference and inhibition in cognition. San Diego:Academic Press, 1995: 207-61.

Parker ES, Weingartner H. Retrograde facilitation of human memoryby drugs. In: Weingartner H, Parker ES, editors. Memoryconsolidation: psychobiology of cognition. Hillsdale (NJ): LawrenceErlbaum, 1984: 231-51.

Pascual-Leone A, Torres F. Plasticity of the sensorimotor cortexrepresentation of the reading finger in Braille readers. Brain 1993;116: 39-52.

Posner MI, Rafal RD. Cognitive theories of attention and therehabilitation of attentional deficits. In: Meier MJ, Benton AL,Diller L, editors. Neuropsychological rehabilitation. Edinburgh:Churchill Livingstone, 1987: 182-201.

Prado de Carvalho L, Zornetzer SF. The involvement of the locuscoeruleus in memory. Behav Neural Biol 1981; 31: 173-186.

Ptito M, Lepore F, Lassonde M, Guillemot JP. Paradoxicalimprovement of residual vision following corpus callosotomy inbrain damaged cats. J Hirnforsch 1993; 34: 43-6.

Purves SJ. Selection of patients for corpus callosum section. In:Spencer SS, Spencer DD, editors. Surgery for epilepsy. Boston:Blackwell Scientific, 1991: 69-86.

Ramachandran VS. Behavioral and magnetoencephalographiccorrelates of plasticity in the adult human brain. [Review]. ProcNatl Acad Sci USA 1993; 90: 10413-20.

Ramachandran VS. Phantom limbs, neglect syndromes, repressedmemories, and Freudian psychology. [Review]. Int Rev Neurobiol1994; 37: 291-333.

Randolph C. Implicit, explicit, and semantic memory functionsin Alzheimer's disease and Huntington's disease. J Clin ExpNeuropsychol 1991; 13:479-94.

Rauschecker JP. Compensatory plasticity and sensory substitutionin the cerebral cortex. [Review]. Trends Neurosci 1995; 18: 36-43.

Rauschecker JP, Kniepert U. Auditory localization behaviour invisually deprived cats. Eur J Neurosci 1994; 6: 149-60.

Rauschecker JP, Korte M. Auditory compensation for early blindnessin cat cerebral cortex. [Review], J Neurosci 1993; 13: 4538-48.

Rauschecker JP, Tian B, Korte M, Egert U. Crossmodal changes inthe somatosensory vibrissa/barrel system of visually deprivedanimals. Proc Natl Acad Sci USA 1992; 89: 5063-7.

Read W, Nenov VI, Halgren E. Role of inhibition in memoryretrieval by hippocampal area CA3. Neurosci Biobehav Rev 1994;18: 55-68.

Rodman HR. Development of inferior temporal cortex in themonkey. [Review]. Cereb Cortex 1994; 4: 484-98.

Ronnberg J. Perceptual compensation in the deaf and blind: mythor reality? In: Dixon RA, Backman L, editors. Compensatingfor psychological Deficits and Declines. Mahwah, New Jersey:Lawrence Erlbaum Associates, 1995: 251-274.

Rosier F, Roder B, Heil M, Hennighausen E. Topographic differencesof slow event-related brain potentials in blind and sighted adulthuman subjects during haptic mental rotation. Brain Res Cogn BrainRes 1993; 1: 145-59.

Serfaty C, Soroker N, Glicksohn J, Sepkuti J, Myslobodsky MS.Does monocular viewing improve target detection in hemispatialneglect? Restorative Neurology and Neuroscience 1995; 9: 77-83.

Shaw C, Aggleton JP. The effects of fornix and medial prefrontallesions on delayed non-matching-to-sample by rats. Behav BrainRes 1993; 54: 91-102.

Sherman SM. Visual fields of cats with cortical and tectal lesions.Science 1974; 185: 355-7.

Sherman SM. The effect of superior colliculus lesions upon thevisual fields of cats with cortical ablations. J Comp Neurol 1977;172: 211-19.

Simon H, Taghzouti K, Le Moal M. Deficits in spatial-memorytasks following lesions of septal dopaminergic terminals in the rat.Behav Brain Res 1986; 19: 7-16.

Smith RS, Doty RL, Burlingame GK, McKeown DA. Smell andtaste function in the visually impaired. Percept Psychophys 1993;54: 649-55.

Soroker N, Cohen T, Baratz C, Glicksohn J, Myslobodsky MS. Isthere a place for ipsilesional eye patching in neglect rehabilitation?Behav Neurol 1994; 7: 159-64.

Spear NE, McKinzie DL, Arnold M. Suggestions from the infantrat about brain dysfunction and memory. In: Delacour J, editor. Thememory system of the brain. Singapore: World Scientific PublishingCompany, 1994: 278-315.

Spitzer M, Braun U, Hermle L, Maier S. Associative semanticnetwork dysfunction in thought-disordered schizophrenic patients:direct evidence from indirect semantic priming. Biol Psychiatry1993; 34: 864-77.

Sprague JM. Interaction of cortex and superior colliculus inmediation of visually guided behavior in the cat. Science 1966;153: 1544-7.

Sprague JM. The role of the superior colliculus in facilitating visualattention and form perception. Proc Natl Acad Sci USA 1991; 88:1286-90.

Squire LR, McKee RD. Declarative and nondeclarative memory inopposition: when prior events influence amnesic patients more thannormal subjects. Mem Cognit 1993; 21: 424-30.

Staubli U, Ivy G, Lynch G. Hippocampal denervation causes rapidforgetting of olfactory information in rats. Proc Natl Acad Sci USA1984; 81: 5885-7.

Townsend J, Courchesne E. Parietal damage and narrow 'spotlight'spatial attention. J Cogn Neurosci 1994; 6: 220-32.

Treffert DA. The idiot savant: a review of the syndrome. [Review].Am J Psychiatry 1988; 145: 563-72.

Trevarthen CB. Double visual learning in split-brain monkeys.Science 1962; 136: 258-9.

Uhl F, Franzen P, Lindinger G, Lang W, Deecke L. On thefunctionality of the visually deprived occipital cortex in early blindpersons. Neurosci Lett 1991; 124: 256-9.

Vuilleumier P, Hester D, Assal G, Regli F. Unilateral spatial neglectrecovery after sequential strokes. Neurology 1996; 46: 184-9.

Page 16: Paradoxical functional facilitation in brain-behaviour ... · Paradoxical functional facilitation in brain-behaviour research ... Situations where damage to intact brain ... paradoxical

1790 N. Kapur

Walker P, Moylan K. The enhanced representation of surface textureconsequent on the loss of sight. Neuropsychologia 1994; 32: 289-97.

Wall PD. The control of neural connections by three physiologicalmechanisms. Prog Brain Res 1987; 71: 239-47.

Wallace SF, Rosenquist AC, Sprague JM. Recovery from corticalblindness mediated by destruction of non-tectotectal fibers in thecommissure of the superior colliculus in the cat. J Comp Neurol1989; 284: 429-50.

Wallace SF, Rosenquist AC, Sprague JM. Ibotenic acid lesions ofthe lateral substantia nigra restore visual orientation behavior in thehemianopic cat. J Comp Neurol 1990; 296: 222-52.

Warrington EK, Weiskrantz L. Further analysis of the prior learningeffect in amnesic patients. Neuropsychologia 1978; 16: 169-77.

Warrington EK, McCarthy RA. The fractionation of retrogradeamnesia. Brain Cogn 1988; 7: 184-200.

Waziri R. Psychosurgery for anxiety and obsessive-compulsivedisorders. In: Noyes R Jr, Roth M, Burrows GD, editors. Handbookof anxiety, Vol. 4. Amsterdam: Elsevier, 1990: 519-35.

Weiller C, Chollet F, Friston KJ, Wise RJS, Frackowiak RSJ.Functional reorganization of the brain in recovery fromstriatocapsular infarction in man. Ann Neurol 1992; 31: 463-72.

et al. Brain stem activation in spontaneous human migraine attacks.NatMed 1995; 1: 658-60.

Weiskrantz L. Some traps and pontifications. In: Weiskrantz L,editor. Analysis of behavioral change. New York: Harper & Row,1968: 415-29.

Werker JF, Gilbert JHV, Humphrey K, Tees RC. Developmentalaspects of cross-language speech perception. Child Dev 1981; 52:349-55.

Winocur G, Olds J. Effects of context manipulation on memoryand reversal learning in rats with hippocampal lesions. J CompPhysiol Psychol 1978; 92: 312-21.

Wray SH, Mijovic-Prelec D, Kosslyn SM. Visual processing inmigraineurs. Brain 1995; 118: 25-35.

Yang TT, Gallen CC, Ramachandran VS, Cobb S, Schwartz BJ,Bloom FE. Noninvasive detection of cerebral plasticity in adulthuman somatosensory cortex. Neuroreport 1994; 5: 7 0 1 ^ .

Young AY, Ellis AW, Flude BM, McWeeny KH, Hay DC. Face-name interference. J Exp Psychol Hum Percept Perform 1986; 12:466-75.

Zola SM, Mahut H. Paradoxical facilitation of object reversallearning after transection of the fornix in monkeys.Neuropsychologia 1973; 11: 271-84.

Weiller C, May A, Limmroth V, Jiiptner M, Kaube H, Schayck R, Received January 24, 1996. Accepted April 19, 1996