Fifteen adult outpatients with high functioning autism or Asperger's Syndrome will be recruited. Fifteen healthy volunteers matched for age and IQ will also be recruited. Participants will undergo an fMRI scan. During the scan they will perform a task. The paradigm is designed to test control inhibition, face processing and language. Therefore, it includes a go-no go test, a face emotion discrimination test and a linguistic sound. The subjects are shown a series of faces with happy or sad expressions in random order. Each face presentation is accompanied by either a standard or variant sound. The sound stimulus is the English phoneme /oe/. The variant sound was created by raising the frequency of the standard sound by 20%. The task consists of four blocks. During each block the subjects are asked to press a button when they see a happy or sad face depending on the block. Each block consists of 96 trials lasting 2 seconds and an initial and final fixation period of 40 sec. For each trial the target image is projected for 500 ms. There is 1250 ms fixation time allowed for response and 250 ms for jitter. Twenty-four of the 96 trials are no go trials and 16/96 are sound oddball trials. The oddball is at least two trials apart (between two variant sounds there are at least two standard sounds). There are 24 faces displayed in total, 12 female and 12 male. Faces switch every 16 trials. Face genders are balanced. The faces were selected with over 80% congruency except for two faces with congruency of over 72%. The correlation between sound (2 conditions: standard versus variant) and target condition (go-no go) is zero. The task was piloted on healthy controls and autistic adults. All subjects performed well.