Digital Biomarkers and AI for Remote Monitoring of Fatigue Progression in Neurological Disorders: Bridging Mechanisms to Clinical Applications.

Background: Digital health technologies are often subject to regulatory requirements. Regulatory auditing processes are complex but necessary to guarantee quality, efficacy and safety of patients. Evolvements such as digitalized clinical trials, and digital biomarkers require a constant adaption of regulatory frameworks. Objective: This review aims to provide an overview on current regulations and standards for digital therapeutics and digital biomarkers, from technical development to market access. Methods: We conducted an unstructured literature review to identify the relevant guidelines, policies and standards for software based digital therapeutics and digital biomarkers. Results: The principal regulations governing software as a medical device are outlined in Chapter 21 of the Code of Federal Regulations by the US Food and Drug Administration, as well as the European Medical Device Regulation 2017/745. Regulatory pathways, such as the DiGA, are in the process of development, particularly for digital therapeutics, which fall within the purview of software as a medical device. Qualification of (digital) biomarkers is typically voluntary but can play a significant role in the development and approval of digital therapeutics. Conclusions: Fragmented, lacking and diverse regulations around digital biomarkers and digital therapeutics highlight the urge to harmonize and foster regulatory frameworks on an international level.

Introduction: Current cognitive assessments suffer from floor/ceiling and practice effects, poor psychometric performance in mild cases, and repeated assessment effects. This study explores the use of digital speech analysis as an alternative tool for determining cognitive impairment. The study specifically focuses on identifying the digital speech biomarkers associated with cognitive impairment and its severity. Methods: We recruited older adults with varying cognitive health. Their speech data, recorded via a wearable microphone during the reading aloud of a standard passage, were processed to derive digital biomarkers such as timing, pitch, and loudness. Cohen’s d effect size highlighted group differences, and correlations were drawn to the Montreal Cognitive Assessment (MoCA). A stepwise approach using a Random Forest model was implemented to distinguish cognitive states using speech data and predict MoCA scores based on highly correlated features. Results: The study comprised 59 participants, with 36 demonstrating cognitive impairment and 23 serving as cognitively intact controls. Among all assessed parameters, similarity, as determined by Dynamic Time Warping (DTW), exhibited the most substantial positive correlation (rho = 0.529, p < 0.001), while timing parameters, specifically the ratio of extra words, revealed the strongest negative correlation (rho = −0.441, p < 0.001) with MoCA scores. Optimal discriminative performance was achieved with a combination of four speech parameters: total pause time, speech-to-pause ratio, similarity via DTW, and intelligibility via DTW. Precision and balanced accuracy scores were found to be 88.1 ± 1.2% and 76.3 ± 1.3%, respectively. Discussion: Our research proposes that reading-derived speech data facilitates the differentiation between cognitively impaired individuals and cognitively intact, age-matched older adults. Specifically, parameters based on timing and similarity within speech data provide an effective gauge of cognitive impairment severity. These results suggest speech analysis as a viable digital biomarker for early detection and monitoring of cognitive impairment, offering novel approaches in dementia care.

BackgroundAnxiety, both generalised anxiety disorder and anxiety symptoms, has been recognised as a risk factor and prodromal symptom of dementia. Digital biomarkers are gaining interest as proxy markers for mental health because they enable passive and continuous data collection, allowing for early detection. However, the association between digital biomarkers and anxiety remains unknown. This systematic review and meta‐analysis aimed to examine the association between digital biomarkers obtained from wrist‐worn wearables and anxiety symptoms in adults.MethodSystematic literature searches were conducted across six databases, including unpublished grey literature. Studies investigating the association between digital biomarkers from wrist‐worn wearables and anxiety were eligible for this review. Effect sizes were combined across studies, for each digital biomarker separately, using random‐effects meta‐analyses whenever possible. Sensitivity analyses were performed to assess whether results differed according to anxiety type (state, trait), age group (young, middle, older), and sex.ResultTwenty‐two articles were eligible. Meta‐analyses were conducted for four sleep metrics: sleep efficiency, wake after sleep onset, total sleep time, and sleep onset latency. Analyses revealed that sleep efficiency (8 studies, Fisher’s z = ‐0.08, 95% confidence interval [CI] = ‐0.15 to ‐0.01, p = 0.0263) and wake after sleep onset (6 studies, Fisher’s z = 0.13, 95% CI = 0.01 to 0.24, p = 0.0291) were associated with anxiety symptoms. In sensitivity analyses, associations persisted older adults (for sleep efficiency only) and trait anxiety. Meta‐analyses could not be conducted for physical activity metrics, however, a qualitative synthesis of the limited number of studies (five studies) revealed inconsistent results.ConclusionWorse sleep efficiency and longer wake after sleep onset were associated with greater anxiety symptoms. However, due to the limited number of studies, the association with physical activity remains unclear, warranting further research. Further, it became apparent that machine learning studies in this area are limited and of variable quality. Given anxiety is a prodromal symptom of dementia, future research focusing on older adults is essential to explore the use of digital biomarkers in the context of dementia risk.

BackgroundDespite growing interest in Behavioural and Psychological Symptoms in Dementia (BPSD) there remains a lack of reliable methods to accurately monitor these disturbances in home settings. This study focuses on two prominent BPSDs ‐ apathy (i.e., a reduction in goal‐directed behaviour) and anhedonia (i.e., a decreased ability to experience and to pursue pleasure). Although distinct, these symptoms often co‐occur in patients with Alzheimer's disease (AD) and behavioural variant of frontotemporal dementia (bvFTD), significantly undermining patients' autonomy and quality of life. Validated and reliable markers of these symptoms are needed to ensure reliable disease monitoring and to inform the development of targeted interventions.MethodsThe ECOCAPTURE@HOME study (Clinicaltrials.gov: NCT04865172) aims to remotely measure behavioural markers of apathy such as daytime activity, quality of sleep, emotional arousal, in everyday life. Here we extend this framework to explore the utility of digital markers in capturing apathy and anhedonia in the home setting in dementia. Spanning two sites, the Paris Brain Institute and the University of Sydney, we aim to recruit 20 AD patient‐caregiver dyads, 20 bvFTD patient‐caregiver dyads and 20 healthy control dyads. Empatica EmbracePlus (Empatica Inc., Boston, MA, USA) wristbands equipped with multiple integrated physiological sensors (e.g., accelerometer, electrodermal activity sensor), alongside targeted behavioural questionnaires and ecological momentary assessments will be implemented. Machine learning methods will be used to explore profiles of apathy and anhedonia at the group and individual patient level.ResultsWe hypothesise that apathy and anhedonia can be reliably identified and tracked using ECOCAPTURE digital biomarkers. Specifically, we propose that apathy markers will exhibit momentary fluctuations contingent on context and environment, whereas anhedonia will remain relatively stable over the course of the day.ConclusionsThis project will contribute novel objective behavioural indices of BPSDs that can be effectively implemented in the home environment. These digital biomarkers will provide essential information to identify windows of opportunity for intervention. Moreover, our approach will enable therapeutic monitoring of targeted interventions supporting patients’ autonomy, alleviating caregivers’ stress, and enabling people with dementia to age in place.

For incurable diseases, such as multiple sclerosis (MS), the prevention of progression and the preservation of quality of life play a crucial role over the entire therapy period. In MS, patients tend to become ill at a younger age and are so variable in terms of their disease course that there is no standard therapy. Therefore, it is necessary to enable a therapy that is as personalized as possible and to respond promptly to any changes, whether with noticeable symptoms or symptomless. Here, measurable parameters of biological processes can be used, which provide good information with regard to prognostic and diagnostic aspects, disease activity and response to therapy, so-called biomarkers Increasing digitalization and the availability of easy-to-use devices and technology also enable healthcare professionals to use a new class of digital biomarkers—digital health technologies—to explain, influence and/or predict health-related outcomes. The technology and devices from which these digital biomarkers stem are quite broad, and range from wearables that collect patients’ activity during digitalized functional tests (e.g., the Multiple Sclerosis Performance Test, dual-tasking performance and speech) to digitalized diagnostic procedures (e.g., optical coherence tomography) and software-supported magnetic resonance imaging evaluation. These technologies offer a timesaving way to collect valuable data on a regular basis over a long period of time, not only once or twice a year during patients’ routine visit at the clinic. Therefore, they lead to real-life data acquisition, closer patient monitoring and thus a patient dataset useful for precision medicine. Despite the great benefit of such increasing digitalization, for now, the path to implementing digital biomarkers is widely unknown or inconsistent. Challenges around validation, infrastructure, evidence generation, consistent data collection and analysis still persist. In this narrative review, we explore existing and future opportunities to capture clinical digital biomarkers in the care of people with MS, which may lead to a digital twin of the patient. To do this, we searched published papers for existing opportunities to capture clinical digital biomarkers for different functional systems in the context of MS, and also gathered perspectives on digital biomarkers under development or already existing as a research approach.

BackgroundDigital biomarkers are gaining interest as proxy markers for mental health, as they enable passive and continuous data collection. However, the association between digital biomarkers of health and anxiety, both generalized anxiety disorder and anxiety symptoms, remains unknown.ObjectiveThis systematic review and meta-analysis examined the association between digital biomarkers of health obtained from wrist-worn wearables and anxiety in adults.MethodsSystematic literature searches were conducted across 6 databases, including unpublished gray literature. The final search was done on September 21, 2025. Cross-sectional or longitudinal studies investigating the association between digital biomarkers from wrist-worn wearables and anxiety were eligible. Studies using inferential statistics or machine learning methods were both eligible. Studies were excluded if participants received diagnoses of neurodegenerative disorders or physical health conditions. Two risk-of-bias tools were used: the National Heart, Lung, and Blood Institute assessment tool for inferential statistical studies, and the modified version of the Quality Assessment of Diagnostic Accuracy Studies-2 for machine learning studies. Whenever possible, effect sizes were combined across studies, for each digital biomarker of health separately, using random-effects meta-analyses. Sensitivity analyses were performed to assess whether results differed according to anxiety type (state or trait) and age group. Otherwise, studies were synthesized narratively.ResultsA total of 44 studies from 42 articles were eligible. Among these, 36 studies used inferential statistical approaches for analysis (21 reporting sleep characteristics, 8 reporting physical activity, 2 reporting heart rate variability, and 5 reporting more than 1 type), and 8 studies used machine learning approaches. Sample size ranged from 17 to 170,320. Meta-analyses on 4 sleep metrics found no associations: sleep efficiency (Fisher z=–0.07, 95% CI –0.14 to 0.002; P=.06; PI –0.19 to 0.05), wake after sleep onset (Fisher z=0.13, 95% CI –0.04 to 0.30; P=.11; PI –0.15 to 0.41), total sleep time (Fisher z=0.009, 95% CI –0.01 to 0.03; P=.28; PI –0.02 to 0.03), and sleep onset latency (Fisher z=0.04, 95% CI –0.07 to 0.15; P=.08; PI –0.19 to 0.27). Qualitative syntheses revealed that lower physical activity levels and higher heart rate were associated with greater anxiety symptoms. Machine learning studies using wrist-worn wearable data alone showed varied performance, with predictive performance improving when wearable data were combined with other data sources.ConclusionsThis is the first review to synthesize evidence from inferential statistical (mostly fair quality) and machine learning studies examining association between wearable-derived digital biomarkers and anxiety. Meta-analyses found no associations between sleep metrics and anxiety. Although based on limited studies, lower physical activity levels and elevated heart rate were associated with greater anxiety symptoms. Digital biomarkers may be more useful when integrated with other data sources (eg, self-report and clinical data) rather than used as stand-alone screening tools.Trial RegistrationPROSPERO CRD42023409995; https://www.crd.york.ac.uk/PROSPERO/view/CRD42023409995

BackgroundRecently developed blood and digital biomarkers may enable the identification of persons with amyloid deposits in the brain. While these biomarkers are promising, most haven’t been evaluated in populations like those screened for clinical treatment trials or in populations with racial and ethnic diversity similar to the US population. The Global Alzheimer’s Platform Foundation® (GAP) completed its biomarker study (Bio‐Hermes) comparing results of blood and digital biomarker tests with brain amyloid PET scans and traditional cognitive tests. The trial enrolled 1,002 participants in three cohorts: Cognitively Normal (CN), Mild Cognitive Impairment (MCI), and Mild AD.MethodWithin 18 months, 17 sites enrolled 1,002 participants (60‐85 y/o) in clinically defined cohorts of: CN (N = 417), MCI (N = 312), and Mild AD (N = 273). Traditional and digital cognitive testing, amyloid PET imaging, biospecimen collection, and speech analytics were performed. A subset of participants received a retinal exam. Blood biomarkers included, Aß 42/40, p‐Tau181/217/231, NfL, GFAP, full genome sequencing, including APOE status, and proteomics. Digital biomarkers included memory recall, executive functions, and drawing‐based and speech elicitation tasks. Blood biomarkers’ relationship to amyloid PET was measured using Spearman’s rank correlation. A ROC curve analysis will assess the sensitivity and specificity of each biomarker and combination of biomarkers compared with amyloid positivity from PET.ResultStatistical models to be competed in April 2023 for presentation.24% of participants were from underrepresented populations (URP) either African American, Hispanic or other. The CN cohort was 61% female, with 21% amyloid PET Positive; the MCI cohort was 54% female with 34% amyloid PET Positive; the Mild AD cohort was 51% female with 62% amyloid PET Positive. Penalized multiple regression will be used to predict amyloid level (continuous) and amyloid positivity status. Candidate predictors include cohorts, blood‐based biomarkers, cognitive assessments (rating scales and digital), retinal measurements, and speech recognition digital biomarkers.ConclusionBio‐Hermes generated a unique, well‐characterized, diverse sample set that determined the utility of several promising biomarkers. Validation of these biomarkers will expedite AD clinical trial enrollment by quickly identifying appropriate participants while reducing the variability and burden of screen failures from high‐cost brain scan procedures.

BackgroundEven if classic neuropsychological tests often have excellent psychometric properties to detect Mild Cognitive Impairment (MCI), they are not suitable for cost‐effective low‐burden screening at scale. Speech‐based digital biomarkers can be deployed in a highly automated fashion. We present the results of an MCI screening algorithm based on a digital Speech Biomarker for Cognition (SB‐C) in the Swedish H70 birth cohort study.MethodWe used a sample from the Swedish H70 Birth Cohort study (N = 404; 356 cognitively healthy (HC), 48 MCI). We automatically extract the SB‐C score and its subscores (executive function, memory, semantic memory, processing speed) from SVF and RAVLT speech recordings using ki:elements’ proprietary speech analysis pipeline including automatic speech recognition and feature extraction. We performed (1) inferential statistics comparing MCI and HC group based on the biomarker scores and (2) built a machine learning model to screen for MCI. For (1) we performed a non‐parametric Kruskal‐Wallis test to compare SB‐C scores of both HC and MCI groups to check for general feasibility. For (2), we trained a support vector machine model with class weights and leave‐one‐out cross validation to classify between MCI and HC using the SB‐C scores as input (overall score and the subscores).ResultThere was a group difference for the SB‐C aggregated cognition score between the groups (HC &gt; MCI; χ2 = 45.9 (1), p &lt;0.001; Figure 1), and also for the subscores (Table 2). To classify between MCI and HC, using a feature selection method, the best model was found for all the five biomarker scores selected with an Area Under Curve of 0.77 (Figure 2), a specificity of 0.77 and a sensitivity of 0.76 (Table 3).ConclusionWe found that a machine learning‐based screening algorithm based on the SB‐C can detect probable MCI patients in representative population sample of older people using a speech biomarker read‐out.

(1) Background: Monitoring of Multiple Sclerosis (MS) with eHealth interventions or digital biomarkers provides added value to the current care path. Evidence in the literature is currently scarce. MS sherpa is an eHealth intervention with digital biomarkers, aimed at monitoring symptom progression and disease activity. To show the added value of digital biomarker–based eHealth interventions to the MS care path, an early Health Technology Assessment (eHTA) was performed, with MS sherpa as an example, to assess the potential impact on treatment switches. (2) Methods: The eHTA was performed according to the Dutch guidelines for health economic evaluations. A decision analytic MS model was used to estimate the costs and benefits of MS standard care with and without use of MS sherpa, expressed in incremental cost-effectiveness ratios (ICERs) from both societal and health care perspectives. The efficacy of MS sherpa on early detection of active disease and the initiation of a treatment switch were modeled for a range of assumed efficacy (5%, 10%, 15%, 20%). (3) Results: From a societal perspective, for the efficacy of 15% or 20%, MS sherpa became dominant, which means cost-saving compared to the standard of care. MS sherpa is cost-effective in the 5% and 10% scenarios (ICERs EUR 14,535 and EUR 4069, respectively). From the health care perspective, all scenarios were cost-effective. Sensitivity analysis showed that increasing the efficacy of MS sherpa in detecting active disease early leading to treatment switches be the most impactful factor in the MS model. (4) Conclusions: The results indicate the potential of eHealth interventions to be cost-effective or even cost-saving in the MS care path. As such, digital biomarker–based eHealth interventions, like MS sherpa, are promising cost-effective solutions in optimizing MS disease management for people with MS, by detecting active disease early and helping neurologists in decisions on treatment switch.

Voice analysis as a digital biomarker: A machine learning approach for automated multiple sclerosis classification.

Objective There is a need for novel biomarkers that can indicate disease state, project disease progression, or assess response to treatment for amyotrophic lateral sclerosis (ALS) and associated neurodegenerative diseases such as frontotemporal dementia (FTD). Digital biomarkers are especially promising as they can be collected non-invasively and at low burden for patients. Speech biomarkers have the potential to objectively measure cognitive, motor as well as respiratory symptoms at low-cost and in a remote fashion using widely available technology such as telephone calls. Methods The PROSA study aims to develop and evaluate low-burden frequent prognostic digital speech biomarkers. The main goal is to create a single, easy-to-perform battery that serves as a valid and reliable proxy for cognitive, respiratory, and motor domains in ALS and FTD. The study will be a multicenter 12-months observational study aiming to include 75 ALS and 75 FTD patients as well as 50 healthy controls and build on three established longitudinal cohorts: DANCER, DESCRIBE-ALS and DESCRIBE-FTD. In addition to the extensive clinical phenotyping in DESCRIBE, PROSA collects a comprehensive speech protocol in fully remote and automated fashion over the telephone at four time points. This longitudinal speech data, together with gold standard measures, will allow advanced speech analysis using artificial intelligence for the development of speech-based phenotypes of ALS and FTD patients measuring cognitive, motor and respiratory symptoms. Conclusion Speech-based phenotypes can be used to develop diagnostic and prognostic models predicting clinical change. Results are expected to have implications for future clinical trial stratification as well as supporting innovative trial designs in ALS and FTD.

Guess What We Can Hear—Novel Voice Biomarkers for the Remote Detection of Disease

Background: Early detection of neurodegenerative disorders remains challenging, particularly in Middle Eastern populations where diagnostic delays and limited biomarker availability impact treatment outcomes. This study aims to build and validate explainable artificial intelligence (AI) models for early prediction of Alzheimer’s disease, Parkinson’s disease (PD) and multiple sclerosis using multimodal data from Saudi Arabian and regional populations. Methods: We constructed deep learning models using data from 12,458 participants (4872 with neurodegenerative disorders and 7586 controls) across 8 centers in Saudi Arabia (KSA) (2018–2024). Models integrated clinical variables, genetic markers, neuroimaging features and digital biomarkers. Model performance was evaluated using AUROC, sensitivity, specificity and prospective validation. Explainability was assessed using SHapley Additive exPlanations values and attention maps. Results: The multimodal ensemble model achieved superior performance (AUROC: 0.92, 95% confidence interval: 0.90–0.94) compared to single-modality models for predicting neurodegenerative disorders 3–5 years before clinical diagnosis. Performance remained robust across different demographic subgroups (age, gender and education levels). Key predictive features included hippocampal volume, specific genetic variants (APOE, LRRK2 and MAPT), subtle gait parameters and digital voice biomarkers. Explainability analyses revealed distinct predictive patterns for each disorder, with neuroimaging features dominating Alzheimer’s predictions and digital biomarkers showing the highest importance for PD. Conclusion: Our explainable AI approach demonstrates high accuracy for early prediction of neurodegenerative disorders in Middle Eastern populations, with potential to reduce diagnostic delays by 2–4 years. The integration of multimodal data and emphasis on model explainability addresses critical barriers to clinical implementation. Our findings highlight the importance of population-specific screening approaches and targeted early interventions for neurodegenerative disorders in Saudi Arabia and similar healthcare settings.

Increasing shortage of neurologists compounded by the global aging of the population have translated into suboptimal care of patients with chronic neurological diseases. While some patients might benefit from expanding telemedicine, monitoring neurological disability via telemedicine is challenging. Smartphone technologies represent an attractive tool for remote, self-administered neurological assessment. To address this need, we have developed a suite of smartphone tests, called neurological functional test suite (NeuFun-TS), designed to replicate traditional neurological examination. The aim of this study was to assess the ability of two NeuFun-TS tests-short walk and foot tapping-to quantify motor functions of lower extremities as assessed by a neurologist. A cohort of 108 multiple sclerosis (MS) patients received a full neurological examination, imaging of the brain, and completed the NeuFun-TS smartphone tests. The neurological exam was digitalized using the NeurEx™ platform, providing calculation of traditional disability scales, as well as quantification of lower extremities-specific disability. We assessed unilateral correlations of 28 digital biomarkers generated from the NeuFun-TS tests with disability and MRI outcomes and developed machine learning models that predict physical disability. Model performance was tested in an independent validation cohort. NeuFun-TS-derived digital biomarkers correlated strongly with traditional outcomes related to gait and lower extremities functions (e.g., Spearman ρ > 0.8). As expected, the correlation with global disability outcomes was weaker, but still highly significant (e.g., ρ 0.46-0.65; p < 0.001 for EDSS). Digital biomarkers also correlated with semi-quantitative imaging outcomes capturing locations that can affect lower extremity functions (e.g., ρ ~ 0.4 for atrophy of medulla). Reliable digital outcomes with high test-retest values showed stronger correlation with disability outcomes. Combining strong, reliable digital features using machine learning resulted in models that outperformed predictive power of best individual digital biomarkers in an independent validation cohort. NeuFun-TS tests provide reliable digital biomarkers of lower extremity motor functions.

BackgroundDigital biomarkers are increasingly being used to support depression assessment by providing objective, continuous, and real-time physiological and behavioral data. However, most existing studies have focused on individual biomarkers, such as sleep or cardiac parameters, while integrative evaluations that capture the multidimensional nature of depression remain limited.ObjectiveThis systematic review evaluated digital biomarkers for depression and synthesized evidence on differences between individuals with depression and controls.MethodsEligible studies included observational or interventional studies examining digital biomarkers for depression with validated outcome measures. We searched major international and Korean databases, including MEDLINE, PsycINFO, CINAHL, IEEE Xplore, Web of Science, Cochrane Library, KISS, RISS, KMbase, and KoreaMed, from inception to December 28, 2025. Risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 tool and the Scottish Intercollegiate Guidelines Network checklist. Meta-analyses were conducted using random-effects models with the Hartung-Knapp-Sidik-Jonkman method, and other outcomes were narratively summarized.ResultsThe search yielded 39,617 records, of which 132 studies involving 57,852 participants met the inclusion criteria. These studies encompassed various digital biomarkers, including sleep, physical activity, cardiac measures, smartphone-derived data, speech, GPS data, and circadian rhythms. A meta-analysis of 22 studies (6947 participants) revealed that individuals with depression had significantly longer sleep onset latency (5 studies; n=292; +4.75 min, 95% CI 2.46-7.04; P=.005; 95% prediction interval [PI] 0.01-10.27) and time in bed (3 studies; n=236; +31.81 min, 95% CI 18.22-45.39; P=.01; 95% PI 2.28-55.16). Physical activity counts were also significantly lower (5 studies; n=462; standardized mean difference −0.71, 95% CI −1.33 to −0.09; P=.03; 95% PI −2.18 to 0.71). Although individuals with depression showed a lower sleep efficiency, higher mean heart rate, and lower SD of normal-to-normal intervals, these differences were not statistically significant. Other digital markers yielded inconsistent results. Overall, these findings indicate that no single digital biomarker sufficiently captures depression-related changes. Instead, the results support the superiority of personalized, multimodal approaches. However, the generalizability of these findings is limited by the lack of standardized data collection protocols and high clinical heterogeneity across studies, as reflected in wide PIs.ConclusionsCertain digital biomarkers, particularly sleep onset latency and physical activity counts, showed consistent average differences between the depression and control groups. However, wide PIs indicate substantial variability across settings, suggesting that no single marker is sufficient for reliable detection. This study advances the field by providing a comprehensive meta-analysis of multidimensional digital biomarkers, establishing a quantitative foundation for objective depression screening and monitoring. These findings support the use of personalized, multimodal digital phenotyping approaches and highlight the need for standardized, clinically interpretable frameworks for real-world depression monitoring.