TL;DR
This study proposes the Hurst exponent (H) from fMRI BOLD signals as a novel biomarker for synaptic excitation-inhibition (E:I) balance in the brain. They found that autistic males, but not females, exhibit reduced H in the medial prefrontal cortex (MPFC), suggesting increased excitation. This finding is linked to differences in social camouflaging abilities between autistic men and women.
What Problem Does This Paper Address?
The prevailing theory for autism's neurobiological underpinnings suggests an imbalance between excitatory and inhibitory (E:I) neural activity. However, evidence for this theory has largely come from studies of rare genetic mutations, leaving a gap in understanding how E:I imbalance might affect the broader autistic population. Furthermore, autism is diagnosed significantly more often in males than females, hinting at sex-specific biological mechanisms. This paper aims to bridge these gaps by developing a non-invasive biomarker for E:I balance and investigating whether E:I imbalance manifests differently in autistic males and females, particularly in key social brain regions like the medial prefrontal cortex (MPFC).
Methods
The researchers employed a multi-pronged approach:
- In Silico Modeling: They developed a biologically plausible recurrent neural network model to simulate local field potentials (LFPs) and blood-oxygen-level-dependent (BOLD) signals. This model allowed them to investigate how changes in the E:I ratio affect spectral properties of neural signals, specifically the 1/f slope and the Hurst exponent (H).
- In Vivo Chemogenetic Manipulations in Mice: To validate their in-silico findings, they used chemogenetics (DREADDs) in mice to experimentally manipulate neuronal excitability in the prefrontal cortex. They then measured BOLD signals to see if H changed as predicted by their models.
- Human fMRI Data Analysis: They applied their findings to human resting-state fMRI (rsfMRI) data from autistic and typically developing (TD) males and females. They analyzed the Hurst exponent (H) in various brain regions, focusing on the MPFC, and correlated these findings with behavioral measures of camouflaging.
Datasets
- Human rsfMRI Data: Data was collected from 136 adult participants (29 TD males, 23 autistic males, 33 TD females, 25 autistic females).
- Mouse rsfMRI Data: Data was acquired from mice undergoing chemogenetic manipulation.
- Gene Expression Data: Data from the Allen Institute Human Brain Atlas and RNA-seq data related to androgen hormone effects were used to investigate the biological underpinnings of sex differences.
Key Insights and Ideas
- Hurst Exponent as an E:I Biomarker: The study establishes the Hurst exponent (H) as a promising neuroimaging biomarker that can reflect changes in synaptic E:I balance, particularly in the MPFC. This is a significant advancement as it offers a non-invasive way to probe E:I mechanisms in humans.
- Sex-Specific E:I Imbalance in Autism: The most striking finding is that E:I imbalance, specifically increased excitation (indicated by reduced H), appears to be localized to the MPFC in autistic males but not in autistic females. This suggests that the neurobiological underpinnings of autism may differ significantly based on sex.
- Androgen Influence on Excitatory Neurons: The research links autism-associated genes in excitatory neurons to genes differentially expressed by androgen hormones. This provides a potential biological mechanism for the observed sex differences in E:I imbalance, suggesting that prenatal androgen exposure might play a role in shaping these differences.
- Camouflaging and MPFC Function: In autistic females, higher MPFC H (indicating more balanced E:I) was associated with better camouflaging abilities. This suggests that intact E:I balance in key social brain regions might enable some autistic women to better mask their social-communicative difficulties.
Key Results
- In Silico: The recurrent neural network model predicted that increased excitation (lower E:I ratio) leads to a decrease in the Hurst exponent (H) in both simulated LFP and BOLD signals.
- In Vivo (Mice): Chemogenetically enhancing excitation in the mouse prefrontal cortex led to a significant decrease in H in BOLD signals, mirroring the in-silico predictions. Silencing both excitatory and inhibitory neurons had no significant effect on H.
- In Humans: A significant sex-by-diagnosis interaction was found in the MPFC (region p32), with autistic males showing reduced H compared to TD males. Autistic females did not show this reduction.
- Behavioral Correlation: In autistic females, increased MPFC H was positively correlated with camouflaging ability (r = 0.60, p=0.001), while no such correlation was found in autistic males (r = -0.10, p=0.63).
Limitations and Future Work
- Sample Size: While the study included a reasonable number of participants, the authors acknowledge that larger sample sizes might be needed to detect subtle sex-by-diagnosis interactions in other brain regions using mass-univariate analysis.
- Normative Ranges: Establishing robust normative ranges for H across different brain regions, sexes, and ages is crucial for better identifying deviations in autistic individuals.
- Developmental Trajectories: The study did not investigate how H changes over development, which is an important area for future research, especially given known age-related changes in H.
- Anesthesia Effects: The mouse studies were conducted under anesthesia, which could potentially influence baseline E:I balance. Future work could explore these manipulations in awake conditions.
- Generalizability: The authors suggest that future work should investigate whether sex-specific E:I imbalance is unique to autism or also present in other neurodevelopmental disorders with a male bias, such as ADHD.
Summary updated