Hemispheric Asymmetries in Neural Resource Allocation in Auditory Language Cortex in Children with Autism and Typically Developing Children: An MEG Investigation

Background: recent neuroimaging investigations of cortical activation patterns in individuals with autistic disorder (AD) have reported local overconnectivity in sensory cortices and related results to neural resource allocation abnormalities in AD. We used Magnetoencephalography (MEG) to investigate the time course of neural activation for speech and nonspeech sounds in typically developing (TD) children and children with AD.

Objectives: to determine if neural activation levels in auditory language cortex (i) were generally higher in children with AD as compared to TD controls; (ii) changed over time with repeated exposure to sounds, and (iii) differed as a function of cerebral hemisphere and stimulus class (tones vs. speech).

Methods: We used a 148 channel Magnes 2500 Whole Head Biomagnetometer System™ (4-D Neuroimaging, San Diego, CA) to record the auditory evoked M100 in 10 TD controls (8 boys, 10y 5mo (1.2)) and 11 children with AD (9 boys, M 11y 0mo (2.0). Stimuli: TONES – 250, 500, 1000, and 2000Hz sinusoids; SPEECH – 4 natural tokens of consonant-vowel syllables. Each stimulus was presented 100 times in an interleaved fashion. Recording consisted of two 4-5m scans of 400 epochs each. Each scan was divided into 4 equal blocks representing the first, second, third, and final 100 epochs and M100 amplitude (femtoTesla, fT)) was measured for each block, stimulus class, hemisphere, and child.

Results: TONES - Neural activation level (grand mean M100 amplitude), was lower in TD (114.0 (8.6). vs. children with AD (120.0fT(9.5)). SPEECH - Overall neural activation level was higher in TD (114.6 (9.5) vs. children with AD (104.0 (5.3). The timecourse of neural activation levels differed sharply for TD vs. children with AD for both TONES and SPEECH. TONES – for TD, M100 amplitude decreased (-17%) in LH (125.0(17.9) to 104.1(34.1)) and increased (+14%) in RH (105.1(11.0) to 120.0(19.1)) from the first to the final block. Children with AD showed the opposite effect, with M100 amplitude that slightly increased (+2%) in LH (108.8(8.7) to 110.5(21.3)) and slightly decreased -5% in RH (126.3(20.4) to 119.5(34.9)). SPEECH – for TD, M100 amplitude increased (+14%) in LH (108.3fT(10.5) to 123.1(12.2)) and increased (10%) in RH (103.9(6.1) to 112.6(10.4)). For children with AD, M100 amplitude decreased (-7%) in LH (103.9(5.0) to 96.7(10.1)) and decreased (-5%) in RH (103.5(10.1) to 98.7(6.0)).

Conclusions: in TD children, we report dynamic hemispheric shifts in resource allocation as a function of time and stimulus class: a rightward shift in neural activation for tones and bilaterally-increasing neural activation for speech. Results provide new data regarding the temporal dynamics of neural resource allocation in sound processing in TD children, and are in good accord with previous neuroimaging reports of hemispheric biases in decoding spectral and temporal features in sounds. We report an opposite effect in children with AD, with a slight leftward shift for tones and bilaterally-decreasing activation for speech. Results provide new data regarding the temporal dynamics of neural resource allocation in sound processing in children with AD, and provide further evidence that neural mechanisms underlying sound processing are organized in a fundamentally different manner in this population.