Frontal EEG and Thermal Asymmetries during the ADOS Predict ASD Severity

Background: Thermoregulatory and other homeostatic abnormalities are often noted in ASD. Such reports have received limited attention. However, they do merit investigation given that temperature impacts nearly all aspects of neuronal functioning, including EEG activity. Although aberrant brain activity has been implicated in the manifestation of social symptoms in ASD, thermoregulatory deficits also have the potential to give rise to cognitive dysfunction, including social deficits. For example, human and rodent studies have demonstrated that body and brain temperature undergo systematic changes during social interaction. Moreover, recent studies employing infrared thermography of the face have found that changes in facial temperature correlate with social-cognitive abilities in non-ASD samples.

Objectives: The objective of this exploratory study is to examine how individual variability in physiological measures, such as thermal and EEG asymmetry, acquired during real-life, person-to-person interaction may potentially serve to index the severity of social deficits in ASD.

Methods: In this ongoing study, non-invasive measures of frontal brain alpha, beta, and theta band hemispheric activity and thermal homeostasis are employed during live administration of the Autism Diagnostic Observation Schedule 2^nd Edition (ADOS-2) via wireless EEG and high-resolution infrared thermography of the face, respectively. Eleven individuals with ASD (3 females) and 10 typically developing individuals (4 females) with an age range of 5-17 years have been tested thus far. EEG is recorded as one continuous epoch spanning the duration of the ADOS-2 session and pre-processed offline to remove ocular and movement artifact via EEGLAB. A left-right hemispheric asymmetry index score for frontal lateral (F8-F7) and frontal medial (FC6-FC5) electrode pairs is then calculated. Infrared thermographic images are also taken at the beginning, middle, and end of the ADOS-2 session. Average temperatures for specific ROIs (i.e., cheeks, eyes, medial canthi, and nose) are then extracted using IRFlash (ICI) and temperature asymmetry (T_L-R) is calculated by subtracting average maximum values in left ROIs from average maximum values in right ROIs.

Results: In a GLM, ADOS scores were predicted by diagnosis (t = -10.2, p < .0001), T_L-R (t = -3.5, p < .006), and diagnosis x T_L-R interaction (t = 2.4, p < .04). ADOS scores and T_L-R were significantly related for ASD (r = -.82, p < .02) but not TD (r = .13, p = .78) participants. EEG alpha and beta asymmetry at F7-F8 electrodes was also found to predict ADOS scores (ps < .003) in interaction with diagnosis (ps < .05). This relationship was, again, uniquely present for ASD participants indicating that left frontal asymmetry is associated with lower comparison scores on the ADOS-2. Finally, significant relationships were found between T_L-R and EEG asymmetry (alpha and theta) for the medial FC5-FC6 (ps < .05) but not the lateral F7-F8 electrodes (ps > .05).

Conclusions: The results of this study reveal an intriguing link between ASD severity, facial temperature asymmetry, and EEG asymmetry. Although the relation between all three variables appears to be complex, our findings thus far highlight the potential utility of physiological indices of social symptom severity in ASD.