Assessing Mental Health and Emotional States by Using Smartphone Photoplethysmography-Based Digital Pulse Waveform Analysis: Cross-Sectional Observational Study.

A feature-selection procedure yielded 7 final pulse-waveform features for psychological association analyses.

• Correlation-based feature selection was applied to reduce multicollinearity among time-, curvature-, and frequency-domain features.
• The 7 selected features were: estimated reflection index, crest time (CT), the third curvature minimum (F/A), the fourth curvature minimum (H/A), the first power spectrum density component, the baseline of Fourier decomposition (V0), and systolic blood pressure.
• Strong within-domain associations were found among time-, curvature-, and frequency-domain features, with comparatively weaker cross-domain correlations.

Depressive symptoms were significantly associated with multiple pulse-waveform features.

• Depressive symptoms (measured by PHQ-9) were significantly related to F/A, V0, and the estimated reflection index.
• Associations were identified using univariate regression with participant-level aggregation and cluster-robust standard errors.
• The sample consisted of 127 students and university employees in Shenzhen, China.

Anxiety symptoms were associated with the curvature-domain feature F/A.

• Anxiety (measured by GAD-7) showed an association with F/A (the third curvature minimum).
• Negative psychological states were primarily associated with time- and curvature-domain features overall.
• Associations were examined using univariate regression with cluster-robust standard errors.

Negative affect was associated with crest time (CT) and F/A, while positive affect measures showed fewer and weaker associations.

• Negative affect (measured by PANAS) was associated with CT and F/A.
• Positive affect measures showed fewer and weaker associations with pulse-waveform features compared to negative affect measures.
• Psychological well-being was assessed using the Satisfaction With Life Scale, Subjective Vitality Scale, and Positive and Negative Affect Schedule.

Emotional valence and arousal were associated with distinct curvature-domain pulse features.

• Valence (assessed via Self-Assessment Manikin) was related to F/A and H/A.
• Arousal was associated with CT and H/A.
• These associations were identified through univariate regression analyses.

Random forest models demonstrated statistically significant but modest predictive performance for negative mental health outcomes.

• Random forest models evaluated multivariate predictive performance using participant-level cross-validation.
• Predictive performance was weaker for positive affect compared to negative mental health outcomes.
• The models were described as showing 'statistically significant but modest predictive performance for negative mental health outcomes.'

Smartphone-derived pulse-waveform features showed acceptable agreement with oximeter-derived features, with time-domain features being more robust.

• Bland-Altman analyses indicated minimal systematic bias for outcomes with significant predictive correlations.
• Comparisons with an oximeter showed significant correlations and acceptable agreement.
• Time-domain features demonstrated greater robustness than reflection-based metrics when compared to oximeter reference signals.
• Participants recorded repeated 4-minute fingertip videos using a custom smartphone app while a fingertip oximeter simultaneously acquired reference pulse signals.

The study recruited 127 participants who performed repeated smartphone-based fingertip pulse recordings under a cross-sectional observational design.

• Participants were students and university employees in Shenzhen, China.
• Smartphone videos were converted into photoplethysmography signals, segmented into beat-to-beat intervals, and summarized into time-, curvature-, and frequency-domain features.
• Features were normalized for heart rate and stature.
• Six validated psychological scales were used: Satisfaction With Life Scale, Subjective Vitality Scale, PANAS, PHQ-9, GAD-7, and Self-Assessment Manikin.