Fractal biomarker of daily activity for women with early onset depression

Depression is a major health issue in adolescence and young adulthood, emphasising the need for early risk identification. Patients with major depressive disorder (MDD) often show disturbed daily rest-activity patterns, but such changes are often confounded by medication intake, comorbidities and disease duration.

In this exploratory analysis, we tested whether there are specific changes in daily rest activity (from wrist-worn actigraphy) in women at the onset of MDD without medication, as compared with age-matched controls.

Participants from the MDD group (age 19-32, 24.73±5.13 (mean±SD), N=15) and control group (age 20-31, 24.89±3.82, N=9) completed ~7 day ambulatory actigraphy recordings, followed by a stringently controlled circadian laboratory protocol to assess endogenous circadian melatonin levels. We analysed the daily rhythm of mean activity levels and non-linear fractal dynamics in eight 3-hour time bin across the 24 hours, correlating these measures with depressive symptom severity and endogenous melatonin levels.

Using approaches from non-linear fractal dynamics, we showed that, compared with healthy controls, women at MDD onset had a higher fractal activity correlation (FAC) during the last hours of sleep, indicating more 'wake-like' patterns (FAC within 0-3 hour before wake: 0.92±0.64 (SD) in MDD vs 0.77±0.18 in controls, p=0.02). The alteration was independent of mean activity level and wake duration but appeared to be associated with depressive symptom severity (p=0.08). Moreover, there was a trend association for altered FAC with endogenous melatonin levels in the MDD group (for onefold increase in melatonin level in the last 3 hours before wake, the FAC increased by 0.33±0.17 (SE), p=0.08).

Pre-wake FAC is elevated in unmedicated women at MDD onset and may serve as a potential biomarker associated with symptom severity and circadian physiology.

These findings provide proof-of-concept evidence that unique fractal motor activity patterns may support early detection of MDD.