Objective Assessment of Hypnotherapy Response Using Functional Near-Infrared Spectroscopy: A Feasibility Study

Background: Functional Neurological Disorder (FND) is characterized by motor and sensory symptoms without identifiable neurological damage. Hypnotherapy is an emerging treatment, but objective markers of neural response are lacking. Functional near-infrared spectroscopy (fNIRS) offers a portable, cost-effective method for monitoring cortical activity in real-world therapeutic settings. Objective: To assess the feasibility of using fNIRS to detect frontal cortical hemodynamic responses during a hypnotherapy session and to classify hypnotic states using machine learning. Methods: Hemodynamic data were collected from one FND patient and five healthy controls using an 8-channel portable fNIRS device during baseline, hypnotic induction, and suggestion phases. fNIRS channels targeted the inferior frontal gyrus (triangular part) and superior frontal gyrus (dorsolateral) bilaterally. Preprocessed fNIRS data were analyzed using a general linear model (GLM) with Gaussian basis functions tiled across 60-second epochs. Beta values were extracted and used as input for machine learning models trained with 10-fold crossvalidation. Model performance was evaluated on the independent FND patient data, and Shapley (SHAP) values were computed to interpret feature importance. Results: The onset of hypnotic suggestion state was associated with decreased HbO and increased HbR concentrations, particularly in the left frontal regions. The fine Gaussian support vector machine (SVM) model achieved the highest test accuracy (66.7%) in classifying onset of hypnotic suggestion vs. baseline state. SHAP analysis revealed that the left superior frontal gyrus was the most important feature consistent with its known role in cognitive control and suggestibility. Conclusion: This study demonstrates the feasibility of using fNIRS to assess brain responses during hypnotherapy. Frontal hemodynamic markers, especially in the left hemisphere, need investigation. Our findings support further development of fNIRS-guided hypnotherapy and neurorehabilitation tools, including virtual reality-based biofeedback systems for FND.