The goal of this neuroimaging study is to investigate how emotional states fluctuate in people with bipolar disorder (BD) compared to healthy controls, and to understand the neural mechanisms driving mood instability. The main questions it aims to answer are: * Can emotional states be decoded from fMRI brain activity using machine learning? * Do individuals with BD show more unstable emotional state trajectories (e.g., high metastability, low fractal scaling) than healthy controls? * Does amplifying positive emotions stabilize brain and emotional dynamics in BD? Researchers will compare individuals with bipolar disorder (BD-I or BD-II, currently depressed or mixed state) to healthy controls without psychiatric history to see whether the BD group shows greater fluctuations in emotional brain activity and whether positive emotion regulation strategies normalize this instability. Participants will: * Complete self-report questionnaires on mood, emotion regulation, anxiety, and daily functioning. * Recall and provide short descriptions of personal positive and negative memories to be used in the MRI task. * Undergo fMRI scanning, including: * Resting-state scans * A Think and Regulate Affective States Task (TReAT) where they recall autobiographical memories, rate emotions, and practice amplifying positive mood. * Structural and diffusion MRI for brain mapping. * Receive physiological monitoring (heart rate, respiration) during scanning. * Complete post-scan surveys on emotional state and task experience. This research will help clarify how the brain supports or disrupts emotional regulation in bipolar disorder and may inform the development of personalized, neurobiologically informed treatments for mood instability.