Seungbeom Seo¹ (), Oh-Sang Kwon¹, Sung-Phil Kim¹; ¹Department of Biomedical Engineering, Ulsan National Institute of Science and Technology, South Korea

Perceptual sensitivity to global motion in random-dot kinematograms (RDK) improves as motion coherence increases, yielding a characteristic psychometric function of the relationship between coherence and accuracy. Neurophysiological studies have identified neural correlates of this effect at multiple stages of visual processing. An early occipito-temporal event-related potential (ERP) component becomes more negative, and a later centro-parietal ERP component becomes more positive at higher coherences. Single-neuron recordings in monkeys likewise reveal neurometric response functions that parallel the animal’s psychometric performance. However, it remains unclear whether non-invasive human electroencephalography (EEG) signals can exhibit a similar quantitative correspondence with behavioral performance. We recorded EEG from human participants performing a motion-direction discrimination task with varying levels of motion coherence in RDK. We analyzed ERP components previously linked to sensory evidence coding and decision-related processing, focusing on a late occipital slow-wave negativity (~550-800ms post-stimulus) and a centro-parietal positivity (~400-700ms). Behavioral accuracy increased monotonically with coherence and was well fit by a psychometric curve. Concurrently, EEG measures varied systematically with coherence, closely mirroring behavioral accuracy. The occipital component became less negative at higher coherence levels, while the centro-parietal positivity increased in amplitude with increasing coherence. Fitting a similar curve to the EEG data yielded neurometric functions that closely resembled the psychometric function and were broadly consistent across participants, indicating a robust correspondence between the brain signals and perceptual performance. ERP indices putatively linked to sensory coding and decision-related processing showed partially dissociable neurometric profiles, suggesting that sensory and decisional stages track motion coherence with distinct dynamics. Our findings demonstrate that EEG-derived neurometric functions can parallel human perceptual performance in a motion-direction discrimination task. In this sense, ERP components provide a quantitative link between EEG signals and perceptual decision-making, offering a useful framework for future investigations in visual cognition research.

Acknowledgements: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (RS-2025-00522357).