Evidence accumulation, not a perceptual threshold, constrains neural and behavioural responses in human collision threat detection

Detection of impending collision is fundamental to many human activities, and is widely assumed to be limited by a ‘looming threshold’. Evidence accumulation models explain decision-making in abstract paradigms, but have not been shown to remain valid for continuously time-varying, ecologically relevant stimuli. Here, we record behavioural and EEG responses in a collision detection task, disprove the conventional looming threshold assumption, and instead provide stringent evidence for a looming accumulation model. Generalising existing model assumptions from stationary to time-varying evidence, we show that our model accounts for previously unexplained observations and full distributions of detection. We replicate a centroparietal pre-decision positivity in scalp potentials, and show that our model explains its onset rather than its buildup, suggesting that neural evidence accumulation is implemented differently, possibly in distinct brain regions, in collision detection compared to previous paradigms. Our findings illustrate the value of connecting basic and applied research on human behaviour.