Mood instability and cognitive impairment are important features of several psychiatric disorders, including bipolar disorder (BD). However, little is known about how they might be related when measured simultaneously and longitudinally. This 10-week prospective, between-subjects experimental study used digital remote monitoring to capture mood and cognition, and instability in these measures. Two groups of participants were recruited: a 'high MDQ' group (n = 37), scoring ≥7 on the Mood Disorder Questionnaire (MDQ) with high daily mood instability, and a 'low MDQ' group (n = 37), scoring ≤5 on the MDQ with low daily mood instability. Five days a week, for 10 weeks, participants completed mood ratings and a contextual cueing task that assessed implicit learning of regularities presented over different temporal scales. Participants with high MDQ scores demonstrated shallower slopes of learning for contextual cueing across days, but not within experimental sessions, compared to individuals with low MDQ scores. Mood instability was not associated with greater instability in reaction times to novel targets. For all participants, learning slopes were significantly steeper for targets repeated more frequently across experimental sessions, e.g. daily compared to alternate days, and within experimental sessions e.g. repeated compared to novel trials, demonstrating successful contextual cueing over varying temporal scales. Our findings highlight that individuals with high levels of mood instability may be less effective at detecting and integrating stimulus regularities across days. Concurrent and high-frequency digital remote monitoring of mood and cognition is feasible and can be used to elucidate relationships between mood instability and cognition.