Age-related neural dynamics revealed by time-domain fNIRS decoding of audiovisual dual-task processing.

Decoding accuracy for age group classification was low during resting state but substantially higher during audiovisual dual-task processing.

• Peak decoding accuracy at rest was 0.651, achieved using kurtosis alone.
• Classification accuracy remained poor across different classifiers during the resting condition.
• During the audiovisual dual-task, classification accuracy reached 0.810.
• The contrast between resting and task conditions demonstrates that age-related neural signatures are best captured under high cognitive load.

A combination of five time-domain fNIRS features achieved the highest classification accuracy for distinguishing younger and older adults during the audiovisual dual-task.

• The five features were variance, peak, time-to-peak, slope, and skewness of oxygenated hemoglobin (HbO) signals.
• Combined, these five features achieved a classification accuracy of 0.810.
• Features were extracted from fNIRS signals collected across temporal, frontal, and parietal cortices.
• A supervised machine learning approach using Support Vector Machine (SVM) was employed for classification.
• Nine time-domain features of HbO were extracted and analyzed in total.

Discriminative optodes were bilaterally distributed but showed functional asymmetry between left and right hemispheres.

• Left temporoparietal regions were primarily involved in auditory processing and multisensory integration.
• Right frontoparietal regions supported attentional control and top-down regulation.
• The spatial distribution was described as 'bilaterally distributed but functionally asymmetric.'
• This asymmetry reflects differential roles of hemispheric regions in the aging brain during complex audiovisual processing.

Time-domain fNIRS features are identified as sensitive markers of cortical reorganization associated with healthy aging.

• The study examined cognitively normal younger and older adults, distinguishing healthy aging from pathological conditions.
• fNIRS data were collected from temporal, frontal, and parietal cortices.
• The findings highlight time-domain fNIRS features as 'sensitive markers of cortical reorganization and potential indicators of healthy brain aging.'
• The results suggest that age-related neural alterations can be reliably decoded from individual brain activity patterns under appropriate cognitive states.

Healthy aging is associated with widespread changes in cortical structure and function, particularly within networks supporting multisensory integration and cognitive control.

• The study focused on cognitively normal aging to isolate age-related neural changes from pathological processes.
• Prior to this study, it remained unclear whether age-related neural alterations could be reliably decoded from individual brain activity patterns.
• The study specifically targeted networks in temporal, frontal, and parietal cortices as relevant to multisensory integration and cognitive control.
• Both resting and complex task conditions were examined to determine under which cognitive state age-related decoding is most effective.