fMRI Evidence for Generalised Arousal as a Substitute for Early Selection in Autism during Conditions of Shifting Visual Spatial Attention

Analysis #2 considers only the task blocks, and not the fixation blocks. We derive from the behavioural response data for each subject an ideal waveform representing the orientation of attention over time. We exclude from this analysis all intervals that contain an incorrect behavioural response or a break in fixation.

We do a regression of this ideal waveform with the fMRI time series, convert the resulting regression coefficients to z-scores, and average these z-scores within each region of interest to produce what we term `activation scores'. These activation scores are then fed into an analysis of variance with factors of diagnostic group, hemisphere, and region of interest.

`fMRI Evidence for Generalised Arousal as a Substitute for Early Selection in Autism during Conditions of Shifting Visual Spatial Attention', Matthew Belmonte, 10 November 2001