As the world's population ages, several neuropsychiatric disorders have received increased attention in recent years. Late-life depression (LLD), one of the most common psychiatric disorders in older adults, seriously affects the daily life and social functioning of the elderly through somatic disorders and cognitive impairment. In patients with LLD, susceptibility to depression is correlated with social role absences, poorer executive function and processing speed, sedentary lifestyle, poor nutrition, irregular sleep, etc. (Szymkowicz et al., 2023). A recent epidemiological meta-analysis of 57, 486 older adults found that the average expected prevalence of LLD is 31.74 % (Zenebe et al., 2021). LLD has been related to a variety of pathogenic mechanisms associated with the underlying age-related neurodegenerative and cerebrovascular processes (Jellinger, 2023). In the elderly population, LLD is a primary mental disorder contributing to global disability, associated with significant impairment in many areas of functioning and a substantial reduction in quality of life. Ultimately, LLD increases the perception of poor health and the resulting demand for health services (Friedrich, 2017).

The diagnosis of LLD is often overlooked, so it is critical to have accurate tools to help identify this disease. The current clinical assessment, identification and effectiveness assessment of LLD are mainly based on history taking, mental status examination and some scale assessments, which may not accurately distinguish LLD from other diseases (Grayson and Thomas, 2013; Kwak et al., 2022). However, compared with adult depression, the clinical features of LLD are often atypical. It is always accompanied by multiple somatization symptoms and progressive cognitive deficits (Zhao et al., 2023). Meanwhile, limited assessment time, medical comorbidity, assessment complexity, and competition for clinical attention make early detection difficult (Alexopoulos, 2005; McClintock et al., 2021; Small, 2009). The lack of objective and stable biological indicators limits research in LLD (Balsamo et al., 2018). The search for clinical biomarkers with high specificity and sensitivity is therefore essential to improve the assessment and prognosis of patients with LLD.

The underlying pathobiology of LLD is still poorly understood, and relevant clinical studies have now been conducted to investigate neuroimaging biomarkers in patients with LLD. LLD vulnerability factors have negative effects on functional brain networks, predisposing networks to a state of fragility and instability (Szymkowicz et al., 2023). The triple network model (Gunning et al., 2021) proposes that depression is related to abnormal functioning of the default mode network (DMN), cognitive control network (CCN), and anterior salience network (ASN). The functional connectivity of the DMN appears to be altered in LLD, and such differences may persist into remission (Wu et al., 2011). The CCN may be particularly important in LLD, which is characterized by executive dysfunction (Dotson et al., 2020; Lockwood et al., 2002). Both structural and functional connectivity of the ASN is reduced in LLD (Yun and Kim, 2021). The above studies are functional magnetic resonance imaging (fMRI) studies of LLD. However, the high cost, high noise, limited task range, limited time resolution, limited possible participants and instability of functional connection (FC) restrict the application of fMRI in psychiatric clinical assistant diagnosis (Pinti et al., 2020). In contrast, functional near-infrared spectroscopy (fNIRS) is an emerging and promising brain imaging technique with favorable ecological validity, ease of use, low cost and high acceptability. fNIRS is a non-invasive and portable optical imaging technique of neural activity and can reach 20–30 mm of the cerebral cortex (Ferrari and Quaresima, 2012). Neural activity can increase regional cerebral blood flow (CBF) through a mechanism called neurovascular coupling (Nippert et al., 2018). Oxyhemoglobin (oxy-Hb) and deoxyhemoglobin (deoxy-Hb) are involved in oxygen transport and glucose catabolism for energy in the cerebral cortex, and the degree of regional brain oxygenation can be reflected some extent by calculating the ratio of oxy-Hb to deoxy-Hb (Fig. 1) (Ferrari and Quaresima, 2012).

The relationship between CBF and cerebral hemodynamics in depressed patients has been reported in some studies (Chithiramohan et al., 2022). However, in elderly patients, the findings may be somewhat limited by the measurement instruments, and this relationship remains to be proven (Aizenstein et al., 2016; Chithiramohan et al., 2022). A previous study demonstrated microvascular oxygenation levels in response to changes in cerebrovascular tone in the elderly (Tan et al., 2016). fNIRS has been used as a method to detect microvascular oxygenation levels in a variety of tissues, including the brain (Soares et al., 2019). Among the unique advantages of fNIRS, it may be a good tool for assessing symptoms of LLD. Therefore, fNIRS may be able to objectively reflect the pathological mechanisms of LLD.

The etiology of LLD may be due to specific regions in the cerebrovascular system, inflammation and cognitive impairment, but the detailed mechanism requires further research (Alexopoulos, 2019). This article aims to review the use of fNIRS in the assessment of depressive symptoms, cognitive functioning and social functioning in patients with LLD, which may help to better understand the pathobiology of LLD and promote its assessment, recovery, resilience, and maintenance of remission.