A Potential Biomarker of Brain Activity in Autism Spectrum Disorders: A Pilot fNIRS Study in Female Preschoolers

Abstract

Autism spectrum disorder (ASD) refers to a neurodevelopmental condition whose detection still remains challenging in young females due to the heterogeneity of the behavioral phenotype and the capacity of camouflage. The availability of quantitative biomarkers to assess brain function may support in the assessment of ASD. Functional Near-infrared Spectroscopy (fNIRS) is a non-invasive and flexible tool that quantifies cortical hemodynamic responses (HDR) that can be easily employed to describe brain activity. Since the study of the visual phenotype is a paradigmatic model to evaluate cerebral processing in many neurodevelopmental conditions, we hypothesized that visually-evoked HDR (vHDR) might represent a potential biomarker in ASD females. We performed a case-control study comparing vHDR in a cohort of high-functioning preschooler females with ASD (fASD) and sex/age matched peers. We demonstrated the feasibility of visual fNIRS measurements in fASD, and the possibility to discriminate between fASD and typical subjects using different signal features, such as the amplitude and lateralization of vHDR. Moreover, the level of response lateralization was correlated to the severity of autistic traits. These results corroborate the cruciality of sensory symptoms in ASD, paving the way for the validation of the fNIRS analytical tool for diagnosis and treatment outcome monitoring in the ASD population.

Keywords: autism; biomarker; fNIRS; female; preschoolers; visual physiology.

Conflict of interest statement

The authors declare that they have no conflict of interest.

Figures

Figure 1

vHDR was reliably detected in TD and fASD participants: average across channels. In all panels, the values in the y-axis are multiplied by 105. (A–C) On the left, the average time course for THb (A), OHb (B), and DHb (C) in response to the radial stimulus (Stim, green line) or the mock stimulus (Mock, grey line) are shown for TD participants. The plots on the right depict the average peak response to Stim vs. Mock across all the subjects. The stimulus-driven signal was significantly higher with respect to the mock condition for all of the metrics (t-test, p < 0.001 for all comparisons). (D–F) The same plots as above for the fASD subjects. The stimulus-evoked vHDR was significantly higher with respect to the mock condition (t-test, p < 0.001 for THb and OHb, p < 0.01 for DHb). The orange lines are the Hb responses to the stimulus in fASD plots. The data are expressed as mean ± sem. ** p < 0.01, *** p < 0.001.

Figure 2

vHDR was reliably detected in TD and fASD participants: best channel. In all panels, values in the y-axis are multiplied by 104. (A–C) On the left, the average time course for THb (A), OHb (B), and DHb (C) in response to the radial stimulus (Stim, green line) or the mock stimulus (Mock, grey line) are shown for TD participants. Plots on the right depict the average peak response to Stim vs. Mock across all of the subjects. The stimulus-driven signal was significantly higher with respect to the mock condition for all the metrics (t-test, p < 0.001 for all comparisons). (D–F) The same plots as above, for fASD subjects. The stimulus-evoked vHDR was also significantly higher with respect to the mock condition in this case (t-test, p < 0.001 for THb and OHb, p < 0.01 for DHb). The orange lines are the Hb responses to the stimulus in fASD plots. The data are expressed as mean ± sem. ** p < 0.01, *** p < 0.001.

Figure 3

Comparison of vHDR between TD and fASD participants. In panels (A–C), values in the y-axis are multiplied by 105; in panels (D–F), values in the y-axis are multiplied by 104. (A–C) Average peak responses across all channels of THb (A), OHb (B) and DHb (C) in TD (green lines) and fASD participants (orange lines). An amplitude analysis revealed that the OHb peak was significantly lower in the fASD population than in controls (t-test, p < 0.05). (D–F) Peak response of THb (D), OHb (E) and DHb (F) in the best channel for TD (green lines) and fASD subjects (orange lines). The OHb peak was significantly lower in the fASD population than in the controls (t-test, p < 0.01). Data are expressed as mean ± sem. * p < 0.05, ** p < 0.01.

Figure 4

Atypical lateralization in visual processing in fASD children. (A) Average time course of OHb response in the left (dashed lines) and the right (solid lines) hemispheres for TD (green) and fASD subjects (orange line). The y-axis values are multiplied by 105. (B) A comparison between the amplitude of the OHb signal in the left and the right hemisphere showed a significant reduction of right responses in the fASD group (a Two-Way mixed model ANOVA, p < 0.01). The y-axis values are multiplied by 105. (C) The laterality index, indicating a lower rightward asymmetry of visual processing in fASD subjects compared to TD children (t-test, p < 0.05). (D) A scatterplot illustrating the distribution of average OHb amplitudes recorded in the right and the left hemisphere in TD and fASD subjects, showing a pronounced rightward bias only in TD children. (E) A linear regression analysis shows a significant correlation in the amplitude of average OHb responses evoked in the right and the left hemispheres for TD participants (Spearman correlation, p < 0.001). This correlation is absent in ASD patients (p = 0.965). The ρ (rho) index indicates the Spearman correlation value. In panels (D,E), the y-axis and x-axis values are multiplied by 105. Levels of statistical significance are shown as ** p < 0.01; *** p < 0.001.

Figure 5

Correlation between the laterality index and clinical scores in fASD children. The ρ (rho) index in each plot indicates the Spearman correlation value. The correlation between Laterality Index (LI) and AQ score (A), VABS score (B), ADOS score (C) and non-verbal IQ score (D). A positive correlation between variables was detected for the AQ score (Spearman correlation, * p < 0.05).