Segel, A. | bioRxiv Science

bioRxiv2026

Using HD-DOT in precision accuracy microgenetic research designs to measure change over time in speech, language, and hearing clinical populations

Documenting change is fundamental to understanding the process of intervention among individuals with communication disorders. This technical report demonstrates the clinical applicability of wearable fNIRS systems and the NeuroDOT processing pipelines for examining within-person cortical dynamics of learning. Using a microgenetic research design and a dense sampling approach, we examined changes in the prefrontal cortical hemodynamic response in an adult female participant who completed the same spoken sentence repetition and auditory fixation tasks across eight sessions. In addition to behavioral accuracy, hemodynamic data were collected with a continuous-wave, multi-channel fNIRS system (NIRSport2) using a prefrontal 20-channel optode montage. Data were processed using NeuroDOT (https://www.nitrc.org/projects/neurodot) (Eggebrecht & Culver, 2019) to: (i) standardize signal quality across the sessions to quantify motion levels and to ensure standardized brain map specificity, (ii) to examine both channel space fluctuations in the hemodynamic response and map changes in cortical activation patterns over the sessions. The signal quality met the predefined criteria for only the first five sessions. Participants repetition accuracy did not improve over the five sessions. Channel-wise analysis revealed that HbO concentration differs significantly over right and left hemisphere channels over the course of the five sessions for the Sentence Repetition task, but not for the Auditory Fixation condition. Brain maps revealed qualitative differences in the pattern of prefrontal cortical activation across the five sessions. Behavioral assessments do not fully capture what occurs during speech repetition tasks, and leveraging neuroimaging can help identify and discriminate between disordered and neurotypical populations.

neuroscience↗

bioRxiv2023

Network level enrichment provides a framework for biological interpretation of machine learning results

Machine learning algorithms are increasingly used to identify brain connectivity biomarkers linked to behavior and clinical outcomes. However, non-standard methodological choices in neuroimaging datasets, especially those with families or twins, have prevented robust machine learning applications. Additionally, prioritizing prediction accuracy over biological interpretability has made it challenging to understand the biological processes behind psychopathology. In this study, we employed a linear support vector regression model to study the relationship between resting-state functional connectivity networks and chronological age using data from the Human Connectome Project. We examined the effect of shared variance from twins and siblings by using cross-validation, either randomly assigning or keeping family members together. We also compared models with and without a Pearson feature filter and utilized a network enrichment approach to identify predictive brain networks. Results indicated that not accounting for shared family variance inflated prediction performance, and the Pearson filter reduced accuracy and reliability. Enhancing biological interpretability was achieved by inverting the machine learning model and applying network-level enrichment on the connectome, while directly using regression coefficients as feature weights led to misleading interpretations. Our findings offer crucial insights for applying machine learning to neuroimaging data, emphasizing the value of network enrichment for comprehensible biological interpretation.

neuroscience↗