Psychobehavioral validity of self-reported symptoms based on spontaneous physical activity

A limitation of self-reports is the presence of recall biases including retrospective distortions of the respondents' experiences. To overcome this concern, ecological momentary assessment (EMA) and day reconstruction method (DRM) have recently been developed. Very recently, we reported the psychobehavioral validity of within-individual temporal variations in momentary depressive mood recorded with EMA by examining co-variant properties with spontaneous physical activity as the external criteria. However, the validity of DRM in this context has not been objectively examined yet. Therefore, in this study, we examined the psychobehavioral validity of DRM by examining temporal associations with spontaneous physical activity and then showed the difference from EMA. Twenty-two healthy undergraduates wore a watch-type computer for two consecutive days and recorded self-reported symptoms (fatigue, depressive mood, and anxious mood) by EMA. They also recorded the symptoms afterward according to the series of behavioral episodes they reconstructed (DRM) about the same days. Physical activity was also obtained using an actigraph built into the watch-type computer. Multilevel analysis showed the significant association between depressive mood recorded with EMA and local statistics (mean activity levels calculated from 60min data length) of physical activity around EMA recordings. However, depressive mood recorded with DRM had no significant association with physical activity. As for fatigue and anxious mood, none of the methods showed significant associations with the local statistics of physical activity. These results imply that depressive mood recorded with EMA would include psychobehavioral information which cannot be captured with DRM.