Frontal Cortex Entropy Derived From Resting-State fNIRS for Brain Age Prediction in Major Depressive Disorder.

OBJECTIVE

This study investigated the potential of frontal brain entropy (BEN) metrics derived from resting-state functional near-infrared spectroscopy (rs-fNIRS) as neurophysiological biomarkers for predicting brain age in patients with major depressive disorder (MDD).

METHODS

Rs-fNIRS data (26 channels) were acquired from the frontal cortex of 49 healthy controls (HCs) and 35 MDD patients. Time-series signals for oxyhemoglobin (HbO), deoxyhemoglobin (HbR), and total hemoglobin (HbT) were extracted. Static BEN was computed as permutation entropy (PE) using the full time series for each channel, while dynamic BEN was derived by calculating PE within consecutive time windows. Support vector regression (SVR) was applied to predict brain age in the HC group using static and dynamic BEN features. The optimal model was then used to predict brain age in the MDD group. Finally, receiver operating characteristic (ROC) curve analysis was conducted to evaluate the discriminative capacity of BEN and brain age metrics for distinguishing MDD from HC.

RESULTS

In HC, the strongest brain age prediction was achieved by dynamic BEN derived from HbR signals (r = -0.62, p < 0.001), while in MDD, the best performance was obtained using dynamic BEN from HbO signals (r = -0.78, p < 0.001). MDD patients exhibited a significantly elevated brain age gap (BAG) compared with HC across all entropy-based features (all p < 0.001). Moreover, both group-discriminative BEN and brain age metrics derived from static and dynamic BEN achieved outstanding diagnostic performance, with several combinations reaching an area under the ROC curve of 1.00 in distinguishing MDD from HC.

CONCLUSIONS

Frontal BEN derived from rs-fNIRS represents a potential neurophysiological biomarker of accelerated brain aging in patients with MDD. Furthermore, brain age estimated from cerebrovascular hemodynamic complexity demonstrates high discriminative ability for identifying MDD. Collectively, these findings suggest that frontal neurovascular complexity metrics may serve as both diagnostic markers for MDD and quantitative indicators of pathological aging progression.