Symposium
Chair: Terri L. Lewis
Discussant: Colin Blakemore
The symposium will consider evidence from a variety of sources onthe nature of experiential influences on the development of perception. Wewill consider evidence from several levels of analysis (behaviour, evokedpotentials, brain imaging), from several ways of manipulating experience,(total deprivation, partial deprivation, biased rearing), and from twomodalities (hearing, vision). Issues to be considered include patterns ofnormal development, mechanisms underlying change, and competitiveinteractions during development within and between modalities. Thediscussion will focus on whether there is congruence in the lessons drawnfrom different methods and levels of analysis and on whether there arecommon principles of development that hold across sensory modalities.
Details of individual items:
paper
In order to comprehend language, we must be able to discriminate,categorize, and recognize the sound patterns of our native language and beable to link critical sound differences to difference in meaning. Researchfrom many laboratories, including ours (e.g. Werker & Tees, 1984; Pegg &Werker, 1997), has shown that human infants are born with broad basedperceptual sensitivities to the sounds of the world's languages, and candiscriminate vowel and consonant differences even if those differences arenot used in their native language. By 10-12 months of life thesesensitivities are modified to reflect exquisite tuning to just thosephonetic differences that are used phonemically in the native language. Itwould seem this effect of experience on perceptual sensitivity wouldoptimally prepare the child for word learning. Yet, when the infant of14-months, who is still a neophyte word learner, engages in the task ofattempting to map those minimally different language-specific sounds on tomeaning, she shows difficulty. Words that are similar (e.g. bih vs. dih)are confused when they are paired with an object, even though the child caneasily discriminate those same words in a simple speech perception task(Stager & Werker, 1997). By 20-months of age, however, when they child hasbecome a more proficient word learner, similar sounding words are no longeras easily confused (Werker, Corocoran, & Stager, 1999). In this talk, I will review both behavioural andelectrophysiological work from our and other laboratories which hasexamined these two reorganizations in perceptual sensitivity. The firstreorganization, the age-related change in nonnative speech discrimination,clearly reveals an effect of listening experience on perceptualorganization. And this effect of experience is evident in neural responsesto speech sounds (Dehaene, in press; Best, et. al., 1999). The secondreorganization, the u shaped dip in distinguishing similar sounding words,reveals more about changes within the child. Here, electrophysiologicalstudies done in collaboration between our lab and the labs of D. Mills andH. Neville show that the child of 14-months picks up and stores lessphonetic detail about words than does the child of 20-months. Theimportance of considering the interrelationship between these two kinds ofmechanisms for change - experientially induced and development - will bediscussed.
paper
Recurrent middle ear disease (MED) in infancy has been associatedwith later difficulties in communication. However, the validity of thislink has been questioned and its mechanisms are unknown. MED is extremelycommon (Hogan et al., BMJ, 1997, 314, 350-353), but is usuallyasymptomatic. Because of this, experimental designs often control poorlyfor previous MED history. Many researchers have dismissed a connectionbetween poor hearing and poor communication because children with resolvedMED are usually audiometrically normal. However, recent research has shownthat more complex, central auditory processing can be impaired in thesechildren. We have examined the effect of MED on binaural hearing usingbinaural unmasking (BU). BU is a measure of target sound detection innoise, based on interaural disparities. None of the subjects had MED at thetime of testing. First, we showed in a retrospective study that childrenwith clinically-treated MED in infancy had reduced BU at 6-12 years.Second, we found in a prospective study that children with persistent MEDin infancy (>50% of the first 5 years) also had significantly reduced BU.In the same study, children with less persistent MED in infancy had normalBU. Third, we found that teenagers who had reduced BU earlier in childhood(6-12 years, first study) had, 6 years later, recovered to normal levels ofBU. Fourth, we found that adults wearing a unilateral earplug, to simulateMED, for one week had reduced BU during plugging and normal BU immediatelyafter unplugging. Finally, we found that animals (ferrets) either rearedwith a unilateral earplug, or experiencing an equivalent period (6-18months) of plugging in adulthood, had reduced BU both during plugging andafter removal of the plug. The BU gradually recovered during a period of2-6 months after unplugging. These results collectively suggest that longperiods (weeks to months) of abnormal binaural input induced by asymmetricor fluctuating conductive hearing loss, such as that occurring in MED,reversibly impair binaural hearing. This would, among other effects, reducethe audibility of target sounds in noisy environments. In addition to reducing sound level, we have found that conductivehearing loss delays the transmission of low frequency sounds through theear. In children with surgically-treated MED in infancy, we have found thatbackward masking (BM) at 8 years is greater than in children without ahistory of MED. BM is a measure of auditory temporal resolution. Greatermasking means poorer separation of sequential sounds. In younger children,we found that BM was much greater than in older children, even without ahistory of MED. Of particular interest is the associated finding thatchildren with specific language impairment (SLI) also have greater BM(Wright et al., Nature, 1997, 387, 176-178). Thus, our results areconsistent with the hypothesis that MED in infancy may contribute to SLI.
paper
Results from ERP and fMRI studies of congenitally deaf andcongenitally blind adults suggest that different subsystems within visionand within audition display different degrees of experience-dependentmodification of cortical organization. Within vision and audition theorganization of systems important in processing peripheral space are mostaltered following unimodal sensory deprivation. Within vision, functionsassociated with the dorsal pathway display marked effects of altered earlyexperience. Hypotheses about the origins of these differential effects ofearly experience are being tested in parallel ERP studies of infants andchildren of different ages.
paper
Children treated for congenital cataract provide a naturalexperiment for studying the effects of experience on visual development. Acataract is an opacity of the natural lens of the eye and, if dense andcentral, prevents all patterned input to the retina. The cataract istreated by surgically removing the natural lens of the eye and replacing itwith an optical correction to restore nearly normal visual input. It iswell known that pattern deprivation during early infancy compromises visualdevelopment. Numerous studies support the general principle that vision isworse if the deprivation was monocular rather than binocular, unlessmonocular deprivation was followed by extensive occlusion of thenon-deprived eye. The usual explanation is that monocular deprivationaffects visual development not only by depriving cortical neurons ofpatterned visual input, but also by uneven competition for corticalconnections between the deprived and non-deprived eyes. Extensive occlusionof the non-deprived eye reduces that uneven competition so that the effectsof monocular deprivation are comparable to those of binocular deprivation.However, most previous studies examined spatial vision, with testsbeginning weeks to years after the deprivation ended. Conversely, bytesting the spatial vision of infants immediately after deprivation endedand by testing an aspect of vision known to be mediated by higher corticalcentres, we have discovered that there are exceptions to this generalprinciple. We used Teller Acuity Cards to measure the grating acuity of 28infants who had been treated for dense central congenital cataracts at 1week to 9 months of age. Immediately after the first optical correction,acuity was no better than that of normal newborns but it improvedsignificantly over the first hour of visual input. Surprisingly, there wereno significant differences in acuity between infants treated for bilateralcataracts (n 12) and those treated for unilateral cataract (n 16),despite the fact that only the unilateral cases had suffered from unevencompetition between the eyes. Nor were any deleterious effects of unevencompetition apparent in measurements of acuity 1 month after treatment.Results from a separate cohort indicate that by 1 year of age, competitiveinteractions have begun to influence recovery from deprivation (Lewis etal., 1995). Thus, immediately after deprivation, recovery appears to bedriven exclusively by visual activity. Only later, after visual activityhas initiated the functional changes that we observed, do competitiveinteractions appear to influence acuity. We also assessed the effects of early visual deprivation on latersensitivity to global motion which, unlike spatial vision, depends onneurons in higher cortical areas, namely extrastriate area MT.Surprisingly, deficits in patients treated for bilateral congenitalcataracts (n 8) were three times larger than those in patients treatedfor unilateral congenital cataract (n 14), even when there had beenlittle patching of the non-deprived eye in unilateral cases. The resultsare consistent with the physiological properties of neurons in area MT andsuggest that beyond the primary visual cortex, competitive interactionsgive way to collaborative interactions that enable a relative sparing ofsome visual functions after monocular deprivation.