ABSTRACT
Background: Early in the Covid pandemic, reports started to emerge of patients with lingering symptoms following recovery from acute infection, so called Long Covid (Miners, 2020;Nath, 2020;Troyer, 2020). Despite the high prevalence of neurological symptoms like brain fog and memory dysfunction (Davis, 2020), most research has relied on surveys or clinical tools typically used to assess declarative memory (Alemanno, 2021;Mendez 2021;Taquet, 2021). To our knowledge, no studies have examined Long Covid patients' ability to learn and consolidate a procedural motor skill. Method(s): We addressed this question in a group of 108 patients with Long Covid and 108 age-and sex-matched controls. Participants performed a well-characterized motor sequence typing task alternating 10-second practice with 10-second rest for 36 trials over 12 minutes (Bonstrup, 2020;Bonstrup, 2019;Buch, 2021). The following day, performance was tested to evaluate overnight consolidation. The behavioral endpoint measure was correct sequence typing speed (Buch, 2021) . Data were fitted to a 3-parameter model (initial performance, maximum performance, learning rate). Simple reaction times (RT) were measured twice: at the beginning and the end of the experimental session. Result(s): On average, patients had experienced 50 weeks of symptoms at time of testing. Long Covid patients' typing speed was slower than healthy controls at the beginning of the session (p=0.00075). While learning rates were comparable across groups (p=0.142), at the end of training Long Covid patients had not reached the same performance level as controls (p=0.046). Overnight consolidation was not statistically different between groups (p=0.58). Fourteen patients were hospitalized during their acute infection;they did not perform differently to patients who were not hospitalized. There were no sex differences when comparing healthy, long covid or combined groups. Pre-task RT was slower in Long Covid patients than in healthy controls (Covid 373+/-131ms, controls 317+/-43ms) but did not predict initial or final typing speed, nor learning rate in either group. Post-task RT was also faster in both groups, marginally more so in patients (Covid 353+/-117ms, controls 314+/-51ms). Conclusion(s): Long Covid patients exhibited slower starting performance. While they learned the skill at about the same pace, they failed to reach the same typing speed as healthy controls by the end of the task. Reaction times, here used as a measure of attention, did not exhibit a direct link with learning metrics.