One of the hallmark features of autism is a deficit in the use of language. This feature, like all of those characteristic of autism, shows tremendous individual variability. Some individuals diagnosed along the autism spectrum show language delays but may eventually acquire language abilities at near-normal levels of performance. Others show much more profound deficits in the use of language and communication. Classically, approximately 50% of individuals affected by autism fail to develop useful speech, and many of these individuals never learn to communicate in any functional way. However, we believe it is possible that, even among those individuals with autism in whom expressive language capabilities are severely limited, an understanding of word meanings may be (to a greater or lesser extent, depending on the individual) preserved. In fact, much anecdotal evidence supports just such a belief; families and therapists working with nonverbal individuals with autism often suspect that these individuals "know" more than they can express.
Because most assessments of language abilities depend on overt behavioral responses (such as verbal output, pointing, signing, or other forms of gesturing), and because these types of behavioral responses are notoriously unreliable in nonverbal, low-functioning individuals with autism, it has traditionally been considered difficult, if not impossible, to achieve an accurate assessment of comprehension or receptive knowledge in this group. Recently, however, several experimental methodologies have been developed and refined to the point that they may prove sensitive enough to provide reliable evidence of such processing, even in the absence of more traditional behavioral responses, such as speech and gesturing.
We propose the use of three such research methods to attempt to detect receptive vocabulary knowledge in low-functioning individuals with autism -- eye movement recording, pupillary dilation monitoring, and event-related brain potentials. Eye movement recording and pupillary dilation monitoring are accomplished by use of an eye-tracking device that records the location of the eyes and the size of the pupils without touching the individual. Event-related potentials involved the use of surface electrodes (applied to the scalp, via a cap or a net) to record electrical signals that indicate activity occurring in the brain. In an initial series of experiments, we propose to demonstrate the ability of these complementary techniques to provide evidence of word knowledge in normal adults (Experiment 1), normally developing children (Experiment 2), and in high-functioning individuals with autism (Experiment 3), all of whom are capable of making reliable behavioral responses to which our measures can be compared. We expect to see different patterns of eye movements, changes in pupillary dilation, and brain electrical activity when participants are presented with known vocabulary items as compared to unknown vocabulary items. Having demonstrated the usefulness of these techniques as measures of receptive knowledge in participants who can make reliable behavioral responses, we will then use these same measures to examine receptive word knowledge in a group in which such responses are not always reliable -- nonverbal, low-functioning individuals with autism (Experiment 4). These participants will not be required to make behavioral responses; instead, we expect to differentiate known from unknown items on the basis of the eye movements, pupillary dilation, and event-related potential evidence alone. Finally, in Experiment 5, we will attempt to use these three measures to assess new word receptive learning in low-functioning individuals with autism. Participants' data will be compared before and after a learning period, in which they are exposed to previously unknown words, to determine whether such exposure results in a pattern of data that more closely resembles that of known words.
The results of these experiments could have far-reaching implications for our understanding of language and thought in autism, and could directly inform the therapeutic techniques we use to attempt to improve the language abilities of individuals with autism. The demonstration that nonverbal individuals with autism know more than they can express would be of extraordinary importance to scientists, therapists, parents, and caregivers alike. For instance, knowing that children can understand language even when they do not speak might support the development of more intensive speech and language therapies using a broader range of modalities to capitalize on an individual's functional preferences or strengths.