Ground Reaction Forces and Joint Moments Predict Metabolic Cost in Physical Performance: Harnessing the Power of Artificial Neural Networks

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

the paper adresses a very important tool and highlights all the important aspects of this investigation.

has a very good writting and presents clearly the main findings.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The core of the study is to explore the accuracy of using artificial neural networks (ANN) to predict metabolic costs from the ground reaction force (GRF) and joint torque data of gait. The study collected data from 270 trials with 20 participants walking on a treadmill and constructed two ANN models: netGRF for GRF data and netMoment for joint torque data. Through training, verification and testing, both models show high prediction accuracy, with netGRF model showing slightly better consistency. The results show that GRF and joint torque data can accurately predict metabolic costs through ANN models, which has important implications for sports science, rehabilitation, assistive technology development, and personalized advancement.In addition, the research and experiments done by the author are detailed, but I would like to make the following suggestions here.

Question 1: Although the ANN model performs well on the training set, the verification of generalization ability is not enough, and whether it can adapt to different populations and conditions needs further verification.

Question 2: The paper does not explicitly discuss the stability and reliability of the proposed artificial neural network (ANN) model in long-term applications. In practical applications, the stability and reliability of the model are indeed very important considerations, as they affect whether the model performs consistently over the course of continuous use and whether it can maintain the accuracy of its predictions.

Question 3: The article does not specify the specific details of the training process: for example, the training algorithm used (such as gradient descent), the learning rate, batch size, regularization techniques, etc. These details are important in order to fully understand the training process and reproduce the findings.

Question 4: Did the study adequately consider the diversity of the experimental design, including different walking speeds, participants of different ages and genders, and different walking surfaces, to ensure that the ANN model is generalised and applicable to a wider range of populations and conditions?

Comments on the Quality of English Language

Minor editing of English language required

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

This article explores the ability of artificial neural networks (ANNs) to predict the metabolic cost of human movement and concludes that both GRF (ground reaction forces) and joint moment data can accurately predict costs through ANN models. However, there are still some issues within the content that require modification or additional explanation by the authors:

1 In the data section, it would be beneficial to supplement with information about the equipment used, the calibration process of the sensors, and the precision of data collection.

2 From the conclusions, it can be seen that both models perform well on the training and validation sets, indicating that GRF and joint moment data can accurately predict metabolic costs through ANN models. However, the article's dataset is quite small, containing only 270 data entries, which poses a risk of overfitting. Could additional data be added, or could publicly available datasets be processed accordingly?

3 When evaluating the models, the article primarily relies on MSE (mean squared error) and R-values for assessment. Could explanations of accuracy, recall, and the F1 score also be included?

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

agree