“Consider a tennis player attempting to hit a topspin in breezy conditions,” explained Dr Pratik Mutha, an associate professor in the IIT’s Department of Cognitive and Brain Sciences and a senior author of the study. “While the player might initially make errors and miss their shots, they eventually succeed by adjusting their actions. A number of mechanisms contribute to this improvement.”

One such process might involve the use of conscious and verbalisable strategies to select a different action, avoiding the recurrence of the error. For instance, a tennis coach might provide specific instructions on positioning the racquet or swinging the arm to execute an accurate topspin despite the breeze.

In addition to such “explicit” learning based on verbal cues, there exist other tacit or “implicit” mechanisms that lead to a more gradual refinement of movements outside conscious awareness. Such implicit processes are insensitive to verbal prompts and lead to more incremental improvements with practice, according to the study.

Dr Mutha’s research group explored whether these distinct learning mechanisms are stimulated by different error sources encountered by the brain during movement. The researchers focused on “sensory prediction errors (SPEs)” and “task performance errors (TPEs)” in their work. SPEs occur when there is a mismatch between the sensory information the brain expects to get and what it actually receives. For instance, the tennis player might anticipate their racquet swinging in a particular way, only to find its trajectory somewhat altered by the breeze. On the other hand, TPEs occur when the intended goal is not achieved, such as the player missing the ball entirely, indicating that the action was unsuccessful.

In a series of experiments, the team used novel perturbations to induce specific errors, which then triggered learning in the participants. They found that different error signals set in motion distinct learning processes in the brain.

“Our study indicated that implicit, subconscious learning is driven only by SPEs. Moreover, contrary to some theories, we also found that TPEs did not activate the implicit system,” said Anushka Oza, a co-first author of the study.

Adarsh Kumar, another co-first author and a former PhD student, added, “TPEs led to a change in peoples’ volitional movement strategy, whereby participants consciously adjusted their actions to avoid repeating the error. Remarkably, when asked to turn off strategy use and ignore the TPE even if it affected task performance, participants demonstrated complete compliance. In contrast, with SPEs, participants compulsively adjusted their movements despite being told to disregard the error.”

The team also noted that when both SPEs and TPEs are present, these two learning mechanisms combine in interesting ways. According to Dr Mutha, automatic adjustments (driven by SPEs) and volitional strategy changes (triggered by TPEs) together set the stage for rapid learning and adaptation to new environments.

“Understanding the source of errors and the neuro-computational mechanisms they stimulate is crucial for optimising learning,” said Apoorva Sharma, a co-author and former IIT student pursuing her PhD at Yale University in the US. “This has clear implications for designing training programmes for athletes and performing artists and designing better rehabilitation programmes for individuals with neurological disabilities, such as stroke.”

Having established the divergence in the learning mechanisms stimulated by different error signals, the team now aims to probe its neural foundations. The researchers plan to use brain stimulation techniques to determine if targeting specific brain regions can selectively enhance or suppress one of these mechanisms. Their ultimate goal is to maximise learning and its retention over time in humans.