Prolonged and dysregulated loneliness states in chronic depression: evidence from an ecological momentary assessment study

Background

Loneliness is a prevalent and clinically relevant risk factor for mental disorders, yet it is predominantly conceptualized as a stable trait assessed via retrospective questionnaires. Little is known about its temporal dynamics in daily life, particularly in chronic depression (CD). CD is strongly associated with early relational adversity and its consequences in adulthood including impaired social functioning and loneliness which contribute to chronic symptom trajectories. We aimed to characterize momentary loneliness dynamics in CD compared to healthy controls (HC) and to examine associations with childhood adversity and trait loneliness trajectories.

Methods

In a prospective ecological momentary assessment (EMA) study, 58 individuals with CD and 42 matched HC completed smartphone-based assessments four times daily during weeks one and five of a five-week study period. EMA measures included momentary loneliness, perceived emotional isolation, and perceived social isolation (visual analogue scales). Derived indices comprised mean levels, within-person variability (RMSSD), temporal persistence (inertia), and duration of high-loneliness episodes. Trait loneliness was assessed using the UCLA Loneliness Scale, and childhood adversity with the Childhood Trauma Questionnaire.

Results

Compared to HC, individuals with CD reported substantially higher mean state loneliness (44.3 vs. 10.9, p < .001), greater variability, and markedly prolonged high-loneliness episodes. High-loneliness states persisted for an average of 3.5 consecutive prompts in CD versus 0.8 in HC, corresponding to approximately one full day in CD versus transient episodes in HC. Similar patterns were observed for perceived social and emotional isolation. Inertia did not differ between groups. Associations between state and trait loneliness diverged by group: in HC, higher mean state loneliness was consistently associated with trait loneliness, whereas this relationship was attenuated in CD. In contrast, perceived emotional isolation showed robust associations with trait loneliness in both groups and emerged as the strongest cross-level marker. In the CD group, higher mean levels (r = .40, p = .002) and longer durations (r = .30, p = .024) of emotional isolation were associated with childhood emotional neglect. Exploratory analyses indicated that higher early emotional isolation was associated with increased odds of worsening trait loneliness over five weeks, although estimates were imprecise due to the small sample.

Discussion

These findings provide novel EMA-based evidence that loneliness in CD is characterized not only by elevated levels but by prolonged and dysregulated temporal patterns, indicating impaired recovery from high-loneliness states. The dissociation between state and trait loneliness in CD, alongside the central role of emotional isolation, suggests altered integration of momentary social experience into global self-perception. Associations with childhood neglect support a developmentally informed model in which early adversity contributes to persistent emotional isolation states. EMA-based assessment of loneliness dynamics may offer a sensitive approach for identifying maladaptive temporal patterns and informing personalized interventions targeting emotional reconnection in CD. However, findings are exploratory and require replication in larger samples.

References

No references