Functional Connectivity Abnormalities Between Primary Auditory Cortex and Broca's Area In Autistic Children: a High Density ERP Study of Phonemic Processing

Background:  The autism spectrum disorders (ASD) are a heterogeneous group of neurodevelopmental disorders that share specific a behavioral phenotype classically described as impairment of social reciprocity and communication, accompanied by repetitive behaviors. Biological models supported by neuroimaging studies are converging on a biological phenotype shared by these disorders, specifically, altered functional connectivity in and among critical cortical networks.  Most of this investigative work has used fc/fMRI data in higher functioning individuals. EEG coherence measures electrocortical synchrony of oscillatory brain rhythms between neural networks, hypothesized as a mechanism for functional connectivity, and can be carried out in behaviorally-challenged young autistic children. Since communication deficits in autism are often associated with receptive and expressive language deficits which may be due in part to a failure of integration and synchronization of phonological information across critical language networks, we sought to test the hypothesis that, during auditory processing of phonemes, autistics show decreased functional connectivity between temporal and frontal language networks. 

Objectives:  To collect cortical auditory evoked potentials in response to phonemic stimuli in autistic and age-matched typical children and to measure phase coherence for both groups between regions activated by phonemes in primary auditory cortex (STG) and defined frontal language networks (Broca’s area).

Methods: Auditory ERP in response to a standard phoneme, //ge//, and a deviant, //be// , were recorded from children (ages 5 to 8 yr) with autism spectrum disorders (n = 6) and age-matched typical controls (n = 8) using high-density array nets and 128 channel EEG. Cortical activation in source space derived from the evoked response to the standard phoneme was mapped onto a standard adult whole head model (inflated cortex) for each subject and average maps of the group data were generated for each groups: ASD and controls.   The coherence spectrum between the activated auditory cortex (LSTG) and Brodmann’s area 45 (Br45) was computed for each subject. Group differences were assessed using a t-test with a Bonferroni correction for multiple comparisons.

Results:  Phase coherence between regions of activation in LSTG and Br45 was reduced in children with ASD in the theta band at 7 Hz  (t-stat 3.9004, p=0.001 before and p=0.05 after bonferroni correction). There was a trend toward significance in the lower theta and delta bands and in the low beta band around 23 Hz which showed a strong trend, but did not withstand Bonferroni correction. Group comparisons showed a trend toward increased activation in LSTG and a decreased activation in the IFG operculum in the ASDs compared to typicals, though these, too, fell short of statistical significance. Additional data will be presented with correlations to autism severity, language level and IQ.

Conclusions:  These results support the view that children with ASD have an impairment of functional connectivity between critical language networks involved in phonemic processing. Impaired phonemic processing across these language networks would be expected to contribute to impairment of receptive and expressive language development.  Our ability to demonstrate this impairment using quantitative EEG underscores the value of this technology in behaviorally challenged young children with autism.