Insights into motor impairment assessment using myographic signals with artificial intelligence: a scoping review

The initial literature search yielded a total of 1,346 studies. After restricting the search period from 2014 to 2024, 1,094 studies remained. Limiting the search to English-language articles further reduced the number to 705. Next, filtering for studies that focused on human subjects resulted in a selection of 449 studies. After removing duplicates, 379 studies were left. Based on the inclusion criteria, title screening reduced this number to 222, and abstract screening further narrowed it down to 183. Finally, after a full-text review, 111 studies were selected for the final analysis (Fig. 1). Out of the selected 111 studies, 64 studies were conducted with only healthy participants [38,39,40, 46, 48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107], and 47 studies included data collected from patients [15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33, 108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135].

We summarized the number and age of participants in the control (healthy) group from studies that included patient groups, as well as the number and age of participants in each patient group. If available, the total number of female and male participants was also summarized. Additionally, we extracted details on the model inputs and outputs, the AI algorithms used, and the performance metrics reported in each study (Tables 1 and 2). Table 1 presents studies in which the AI models performed classification tasks, while Table 2 summarizes those involving regression tasks. In the following sections, we provide a detailed review of the selected studies with respect to each of the three scopes and describe the results of the analysis proposed.

Table 1 Summary of studies employing AI models for classification tasksTable 2 Summary of studies employing AI models for regression tasks3.1 Target application

Various measurement modalities were used to measure myographic signals (Fig. 3), including surface electromyography (sEMG), intramuscular EMG (iEMG), high-density sEMG (HD-sEMG), sonomyography (SMG), mechanomyography (MMG), force myography (FMG), and optomyography (OMG). Among these modalities, EMG was the most frequently used measurement modality in all the reviewed studies (Fig. 3A), which accounts for 79.5% including sEMG (67.2%) [15,16,17,18,19,20, 22,23,24,25, 28, 30, 38,39,40, 48, 50,51,52,53,54,55,56,57,

Comments (0)

No login
gif