Diagnostic accuracy is an unmet need for major depressive disorder (MDD) and major depressive episode (MDE) in bipolar disorder. Very limited research has evaluated bipolar disorder/MDE and MDD using ecological momentary assessment (EMA) time-series data.

We aimed to examine differentiating phenomenological characteristics in positive affect dynamics, and temporal relationships with pleasure towards current activity and meaning in life (MIL), among MDD, MDE/bipolar disorder and healthy controls using EMA.

Participants (N = 88, mean age 28.7 years, 69% female), including individuals with MDD (n = 29) and MDE/bipolar disorder (n = 29) and healthy controls (n = 30), were assessed for positive affect, pleasure and MIL 5 times daily over a 2-week period. Multilevel modelling analysis was conducted, with estimation of first-order autoregressive model structure and time-lagged relationship between pleasure and positive affect.

From 4632 EMA observations, positive affect dynamics (inertia, variability and instability) did not differ significantly across groups (all P > 0.05). Although all groups demonstrated a bidirectional relationship between positive affect and pleasure, for MDE/bipolar disorder, both pleasuret − 1 (β = −0.11, t[51.09] = −2.31, P = 0.025) and positive affectt − 1 (β = −0.13, t[56.54] = −2.30, P = 0.025) predicted subsequent MIL less significantly than for MDD and healthy controls.

Individuals with MDE/bipolar disorder, but not MDD, had less self-reported MIL from positive affect and pleasure. There is little evidence that emotional experience alone characterises the pathophysiology between MDD and MDE/bipolar disorder; such investigation may be limited by within-group heterogeneity. Our findings provide a new perspective on using a time-series approach beyond bimodal measures in EMA to differentiate bipolar disorder/MDE and MDD.