Abnormal behavior of following peers in an online game indicates bipolar disorder and manic/hypomanic episodes

Early detection of bipolar disorder (BD) is crucial for its ultimate control and prevention. We aim to detect people with BD and their manic/hypomanic episodes through their abnormal behaviors of following peers in Massively Multiplayer Online Game (MMOG) logs. The total participants consisted of 198,605 users of Pigg Party: a popular MMOG in Japan. Their behaviors, including gacha (an online capsule toy game) and purchase in MMOG, were recorded in milli seconds for one month. Of the total participants, 291 responded to the mood disorder questionnaire. Among the questionnaire respondents, 20 were judged as BD group and the other 271 as non-BD group. According to the anomaly scores of the behaviors of following peers, 13,650 of the total participants were estimated as BD group and the other 184,955 participants as non-BD group. Among the total participants, the BD group had significantly more gacha and purchase behaviors than the non-BD group. Further, among this BD group, these behaviors were significantly more prevalent especially during manic/hypomanic episodes. Our findings indicate that behaviors of following peers in MMOG logs is potentially useful for detecting BD and manic/hypomanic episodes.