Researcher(s)
- Neehal Pathak, Human Physiology, University of Delaware
Faculty Mentor(s)
- Fabrizio Sergi, Biomedical Engineering, University of Delaware
Abstract
The ability of the central nervous system to modulate fast motor responses based on task goals is a key aspect of motor control. The stretch reflex is an important motor response and consists of three components: the short-latency response (SLR), the long-latency response (LLR), and the voluntary response. While SLRs are highly stereotypical and mostly determined by spinal circuits, the LLR, which incorporates spinal and supraspinal pathways that include the corticospinal and the reticulospinal tract (RST), can be tuned based on task goals.
Understanding this modulation and its effects is important to understand the stretch reflex and different motor pathways contributions to motor control in physiological and pathological conditions. Ten participants were subjected to robot-applied wrist extension via the Dual-Motor-Stretch-Wrist under three instructions: “Yield” to the perturbation, “Resist” it, or “Slow”, which is a slow perturbation. Electromyography (EMG) was recorded from the FCR and ECU to measure muscle activity via EMG sensors. EMG normalization is a method used to compare raw EMG signals across perturbations and different participants which was previously normalized to the average EMG 200 ms prior to the perturbation. Multiple EMG normalization methods have been described in the literature, but the effect of different normalization methods has not been fully examined.
Three different normalization methods were implemented on the same dataset: 1) Normalizing to the peak of the EMG signal measured during the resist condition, 2) to the peak of EMG signal measured during the yield condition, and 3) to the average peak EMG signal measured across all conditions. Our results demonstrated that the ‘Resist’ condition elicited the largest LLR response, which was significantly greater than both the ‘Yield’ and ‘Slow’ conditions. Conversely, the ‘Slow’ condition produced the lowest LLR amplitude, indicating a clear and effective suppression of the reflex pathway. The ‘Yield’ condition resulted in an intermediate level of reflex activity. The ‘Yield’ condition had the highest p value as compared to the ‘Slow’ and ‘Resist’ conditions. As such, we conclude that the EMG normalization methods considered in this study are equivalent for the ‘Slow’ and ‘Resist’ conditions but is slightly less for the ‘Yield’ condition.