Dissecting errors made in response to externally- and internally- driven visual tasks in the common marmosets and humans

Various visual paradigms in oculomotor research have been used for studying the neural processes of eye movement, cognitive control, attention and neurological disorders. However, we usually analyse data collected from humans or over-trained non-human primates (NHPs), focusing only on successful trials, whereas error trials are usually excluded. These errors may repetitively show up in diseases, but their interpretation would be difficult due to the absence of records taken from healthy controls. In the current study, we aimed to analyse both correctly and incorrectly performed trials in both marmosets and humans. We trained marmosets to perform the gap saccade task and the oculomotor delayed response task. We also collected data from human subjects who performed identical tasks. We categorised error trials into three different groups, based on the time when an incorrect response occurred. We also interpreted possible causes by analysing saccade reaction time, saccade landing position and task history. Despite the rareness of human error, we found that marmosets and humans showed remarkably similar behaviour in error and success. We also found that successful saccades in the gap saccade task had always the highest peak velocity in both species, reflecting faster sensorimotor processing for correct responses. Our results suggest that marmosets and humans might share similar neural processing for successful and unsuccessful oculomotor behaviour, making them a suitable model for studying human behaviour. More importantly, analysing error trials in sync with successful ones will provide further insights into the cognitive and sensorimotor processes. NEW AND NOTEWORTHYThis is the first detailed report focusing on analysing both error and successful trials in oculomotor tasks. We proposed nomenclatures and a generalizable way of grouping and analysing error trials. Our results also indicate that marmosets can be a promising experimental candidate for oculomotor research because they replicate the saccade properties of error and success seen in humans. This will help set the baseline measurements to study brain disorders using NHP and understand the neural mechanisms from a different perspective.