How distributed features bind into unified percepts — temporal correlation hypothesis
The binding problem asks how the brain integrates features processed in separate cortical regions (color in V4, motion in MT, shape in IT) into unified object representations. The temporal correlation hypothesis (von der Malsburg 1981; Singer & Gray 1995) proposes that neurons representing features of the same object synchronize their firing at gamma frequencies (30–80 Hz), while neurons representing different objects fire at different phases. This simulation shows feature neurons as oscillators: neurons bound to the same object synchronize (Kuramoto coupling); neurons from different objects maintain phase separation. The synchrony index R measures coherence within each object's neural assembly.