Identification of neuro cognitive markers of sarcopenia disease using functional near infrared spectroscopy and artificial intelligence approaches

Sarcopenia, defined as a decline in muscle mass, strength and function associated with aging, is a significant disease that severely impacts quality of life. Recent studies suggest that sarcopenia is not limited to physical deterioration alone but may also negatively affect neurological functions. In this thesis, the cortical hemodynamic effects of sarcopenia during motor and cognitive tasks and their relationship with clinical measurements were investigated using functional near-infrared spectroscopy (fNIRS). Additionally, the efficacy of distinguishing sarcopenic individuals from healthy individuals using machine learning models utilizing functional connectivity features derived from fNIRS data was assessed. A total of 68 participants (38 healthy, 30 sarcopenic) were included in the study. Participants underwent hand grip, N-Back and Oddball tasks to collect fNIRS data, and electromyography (EMG) and hand dynamometer data were additionally recorded during the grip strength task. Statistical analysis of the data was performed using t-tests, repeated measures ANOVA, and correlation analysis. The results revealed that sarcopenic individuals exhibited significantly lower performance on motor tasks, as indicated by EMG and dynamometer measurements. Furthermore, during grip, N-Back and Oddball tasks, hemodynamic responses in the precentral gyrus (PcG), middle frontal gyrus (MFG) and superior frontal gyrus (SFG) regions differed between the sarcopenic and control groups. Correlation analyses between motor and cognitive tasks supported the notion of shared cortical representations for motor and cognitive functions in sarcopenic individuals. Classification using machine learning yielded high accuracy (%80-86) and area under the curve (AUC) values (0.86-0.95) in distinguishing sarcopenic and healthy individuals. These findings indicate that fNIRS-based neurological biomarkers hold significant potential for diagnosing and clinically assessing sarcopenia.