<![CDATA[Bipolar disorder is a mental health condition that causes extremely dramatic mood swings, from very high (mania) to very low (depression). Physically, mood changes can be identified from variations in sound frequency. However, early detection of individuals suspected of having bipolar disorder remains a challenge due to the limited availability of tools and healthcare facilities. This research aims to design a bipolar disorder detection tool based on anomalies in the frequency of conversational sound waves using the MAX9814 sensor as a preventive measure for bipolar disorder, with the novelty of the research being its connection to IoT, enabling real-time monitoring of bipolar patients' emotional conditions. The system design uses Arduino Nano as the data processor, and ESP8266 module for IoT connectivity, with testing method conducted on ten respondents with five variations of emotion combinations. The testing method was performed on 10 people (5 men and 5 women) aged 20–40 years, who recited sentences combining variations of depression and mania. The research results show that the device is able to detect changes in voice frequency with an average error of 5% and an accuracy of 95%. The range of sound frequencies indicative of bipolar disorder is 190–355 Hz, with the following patterns: anger < 320 Hz, happiness < 300 Hz, sadness and fear > 240 Hz, especially if negative sounds are the highest. In individuals with bipolar disorder, voice frequencies can change suddenly by more than ±50 Hz within a single sentence. These findings prove that the developed tool has high sensitivity in detecting emotional changes based on voice frequency and can be used as a supporting instrument in the early detection of bipolar disorder efficiently and sustainably.]]>
Copyrights © 2025
