The High Frequency Brain Response to Illusory Contour in Boys with Autism: The Missing Processing Stage?

Background: The atypical visual perception is frequently reported in children with autism Spectrum Disorder (ASD) and has been hypothetically related to the decreased integration of perceptual information (Frith, 1989) or to over-functioning of the lower-order visual areas (Mottron, 2006).  We used the high-frequency brain oscillatory response during illusory contour perception as a model for studying the neural dynamic underlying automatic contextual integration of local image feature.  

Objectives: This study was designed to determine whether the illusory contour (IC) processing could be reflected in phase-locked gamma and beta band oscillations in typically developing boys (TDB), and whether visual processing abnormalities in boys with ASD would appear in their abnormal gamma band responses to illusory Kanisza figure. 

Methods: EEG data were obtained in 23 ASD boys aged 4-7 years and 23 age-matched TDB.  The stimuli consists of four symmetrical black inducer discs that were arrange in such a way as to either produce a Kanizsa square illusion (Illusory Contour: 134 trials) or not (Control stimulus: 134 trials).  To maintain children’s attention on the computer screen the test stimuli were interspersed with the short animation movies.  The stimulus duration was 500 ms and inter-stimulus interval varied between 500 ms and 1000 ms.  To quantify the power of gamma and beta bands evoked response to the stimulus onset, the complex wavelet transformed data for the respective frequency bands from each single trial were averaged. From this complex average the square modules were taken as a measure of the power of the phase-locked (PL) response. Normalized time courses of PL gamma and beta band responses were analyzed. 

Results: We found phase-locked beta and gamma-band activity to be a very strong component of brain response to illusory contour as compared to non-illusory stimulus in both autism and control children. The high-frequency response to IC was, however, qualitatively different in TDB and ASD. In TDB the IC effect has been observed during two time windows of stimulus processing (40-120 and 120-270 ms).  As compared to the control stimulus the IC evoked higher gamma band response in the later time window (120-270 ms, direct IC effect), but lower gamma response in the earlier window (40-120 ms, inverted IC effect).  The ASD demonstrated abnormally protracted inverted IC effect, but were lacking the later direct IC effect.  

Conclusions: Our results demonstrate that illusory contour response does present in PL gamma oscillation in healthy children and contains two different phases. The earlier phase of inverted IC effect, in the context of literature (Ramsden et al, 2001), may be regarded as a correlate of sensory coding of stimulus features and signaling of illusory lines presence in the contour . The second phase of direct IC effect is most probably associated with perceptual grouping processes related to excitatory feed-back signal from higher-order visual areas subserving shape discrimination (Lamme and Roelfsma, 2000). Children with autism may rely more heavily on the lower-order processing in the primary visual area and are lacking the later stage related to higher-order contour integration process.