Pain Level and Pain-Related Behaviour Classification Using GRU-Based Sparsely-Connected RNNs

Dehshibi, Mohammad Mahdi; Olugbade, Temitayo; Díaz-de-María, Fernando; Bianchi-Berthouze, Nadia; Tajadura‐Jiménez, Ana

doi:10.1109/jstsp.2023.3262358

Cited by 5 publications

(3 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is a highly complex activity that can involve very different strategies, especially to deal with worry about the activity or low confidence in the ability to complete it, and levels of both of worry and confidence can change during the activity as found in our dataset. In fact, findings in the literature highlight differences in strategies used by different people with chronic pain for the same movements, compared with healthy people who show less variability [63]- [65]. This could explain the poorer performance with inclusion of data from healthy people.…”

Section: B Activity Recognition For Movements Of People With Chronic ...mentioning

confidence: 99%

The EmoPain@Home Dataset: Capturing Pain Level and Activity Recognition for People with Chronic Pain in Their Homes

Olugbade¹,

Buono²,

Potapov³

et al. 2023

Preprint

View full text Add to dashboard Cite

<p>Chronic pain is a prevalent condition where fear of movement and pain interferes with everyday functioning. Yet, there is no open body movement dataset for people with chronic pain in everyday settings. Our EmoPain@Home dataset addresses this with capture from people with and without chronic pain in their homes, while they performed their routine activities. The data includes labels for pain, worry, and movement confidence continuously recorded for activity instances for the people with chronic pain. We explored two-level pain detection based on this dataset and obtained 0.62 mean F1 score. However, extension of the dataset led to deterioration in performance confirming high variability in pain expressions for real world settings. We investigated activity recognition for this setting as a first step in exploring the use of the activity label as contextual information for improving pain level classification performance. We obtained mean F1 score of 0.43 for 9 activity types, highlighting its feasibility. Further exploration, however, showed that data from healthy people cannot be easily leveraged for improving performance because worry and low confidence alter activity strategies for people with chronic pain. Our dataset and findings lay critical groundwork for automatic assessment of pain experience and behaviour in the wild. </p>

show abstract

Section: B Activity Recognition For Movements Of People With Chronic ...mentioning

confidence: 99%

The EmoPain@Home Dataset: Capturing Pain Level and Activity Recognition for People with Chronic Pain in Their Homes

Olugbade¹,

Buono²,

Potapov³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…This paper aims to provide a brief discussion of the differences between hormonal computing and conventional methods like deep learning in the context of pain level and pain-related behavior analysis in gait studies. To illustrate this comparison, we refer to a recent work, Dehshibi et al (2023) , where a GRU-based sparsely connected recurrent neural network (RNN) architecture was utilized for pain classification. In recent years, the study of pain and its implications in various fields has gained significant attention.…”

Section: Use Casesmentioning

confidence: 99%

Hormonal computing: a conceptual approach

Vallverdú,

Talanov,

Leukhin

et al. 2023

Front. Chem.

View full text Add to dashboard Cite

This paper provides a conceptual roadmap for the use of hormonal bioinspired models in a broad range of AI, neuroengineering, or computational systems. The functional signaling nature of hormones provides an example of a reliable multidimensional information management system that can solve parallel multitasks. Two existing examples of hormonal computing bioinspired possibilities are shortly reviewed, and two novel approaches are introduced, with a special emphasis on what researchers propose as hormonal computing for neurorehabilitation in patients with complete spinal cord injuries. They extend the use of epidural electrical stimulation (EES) by applying sequential stimulations to limbs through prostheses. The prostheses include various limb models and are connected to a neurostimulation bus called the central pattern generator (CPG). The CPG bus utilizes hormonal computing principles to coordinate the stimulation of the spinal cord and muscles.

show abstract

“…High accuracy was obtained by a Pain Level Assessment with Anomaly-detection based Network or PLAAN in short, in a proposed 'lightweight' LSTM-DNN network(Li et al 2021). This year, a sparsely-connected recurrent neural networks (s-RNNs) ensemble with gated recurrent unit (GRU) that incorporates multiple auto encoders using a shared training framework produced high accuracy(Dehshibi et al 2023). Using the BioVid Heat Pain Dataset, another framework proposed an acute nociceptive pain recognition system using physiological signals and a hybrid deep learning network combining shallow CNN and LSTM network for pain intensity classification and achieved an accuracy of 91.43%(Subramaniam et al 2020).…”

mentioning

confidence: 99%

Facial Expressions as Behavioral Indicators for Assessing Pain using Machine Learning Models

Nath

2023

J Stud Res

View full text Add to dashboard Cite

Pain is a symptom of a condition or disease. Pain experienced in the body is verbally reported to a health care giver. Currently there is no objective way to measure physical pain or discomfort one may be feeling. And so consequently, there is no way for caregivers to adequately assess patients in pain who cannot verbalize it, such as non-verbal, adult patients and young children. Facial expressions may be used as a behavioral indicator for evidence of pain which can then be used to communicate a patient's distress and pain severity. These facial expressions can be recognized through jaw clenching, eyebrow raising, and eye squinting. Machine learning with vision based algorithms may differentiate these behavioral face-indicators and assess the pain levels of nonverbal patients. There have emerged many vision based methods for predicting pain from face images. This review summarizes the development of pain recognition from facial expressive images or videos, datasets available for research, an overview of vision based methods using conventional and deep learning, the current challenges and limitations, and scope for improvement in future.

show abstract

Pain Level and Pain-Related Behaviour Classification Using GRU-Based Sparsely-Connected RNNs

Cited by 5 publications

References 55 publications

The EmoPain@Home Dataset: Capturing Pain Level and Activity Recognition for People with Chronic Pain in Their Homes

The EmoPain@Home Dataset: Capturing Pain Level and Activity Recognition for People with Chronic Pain in Their Homes

Hormonal computing: a conceptual approach

Facial Expressions as Behavioral Indicators for Assessing Pain using Machine Learning Models

Contact Info

Product

Resources

About