Comparison of Deep Learning Algorithms in Predicting Expert Assessments of Pain Scores during Surgical Operations Using Analgesia Nociception Index

There are many surgical operations performed daily in operation rooms worldwide. Adequate anesthesia is needed during an operation. Besides hypnosis, adequate analgesia is critical to prevent autonomic reactions. Clinical experience and vital signs are usually used to adjust the dosage of analgesics. Analgesia nociception index (ANI), which ranges from 0 to 100, is derived from heart rate variability (HRV) via electrocardiogram (ECG) signals, for pain evaluation in a non-invasive manner. It represents parasympathetic activity. In this study, we compared the performance of multilayer perceptron (MLP) and long short-term memory (LSTM) algorithms in predicting expert assessment of pain score (EAPS) based on patient′s HRV during surgery. The objective of this study was to analyze how deep learning models differed from the medical doctors′ predictions of EAPS. As the input and output features of the deep learning models, the opposites of ANI and EAPS were used. This study included 80 patients who underwent operations at National Taiwan University Hospital. Using MLP and LSTM, a holdout method was first applied to 60 training patients, 10 validation patients, and 10 testing patients. As compared to the LSTM model, which had a testing mean absolute error (MAE) of 2.633 ± 0.542, the MLP model had a testing MAE of 2.490 ± 0.522, with a more appropriate shape of its prediction curves. The model based on MLP was selected as the best. Using MLP, a seven-fold cross validation method was then applied. The first fold had the lowest testing MAE of 2.460 ± 0.634, while the overall MAE for the seven-fold cross validation method was 2.848 ± 0.308. In conclusion, HRV analysis using MLP algorithm had a good correlation with EAPS; therefore, it can play role as a continuous monitor to predict intraoperative pain levels, to assist physicians in adjusting analgesic agent dosage. Further studies may consider obtaining more input features, such as photoplethysmography (PPG) and other kinds of continuous variable, to improve the prediction performance.

[1]  J. Shieh,et al.  Pain and Stress Detection Using Wearable Sensors and Devices—A Review , 2021, Sensors.

[2]  K. Alizadeh,et al.  A Novel Classification Method Based on Multilayer Perceptron-Artificial Neural Network Technique for Diagnosis of Chronic Kidney Disease , 2020, Annals of Military and Health Sciences Research.

[3]  Hyuk-Jae Chang,et al.  Predicting Cardiac Arrest and Respiratory Failure Using Feasible Artificial Intelligence with Simple Trajectories of Patient Data , 2019, Journal of clinical medicine.

[4]  Ruslan Abdullayev,et al.  Analgesia Nociception Index: assessment of acute postoperative pain , 2019, Revista brasileira de anestesiologia.

[5]  B. Çelik,et al.  Analgesia Nociception Index: Heart Rate Variability Analysis of Emotional Status , 2019, Cureus.

[6]  Alexander J. Casson,et al.  Human activity recognition from inertial sensor time-series using batch normalized deep LSTM recurrent networks , 2018, 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[7]  Jane C. Deng,et al.  Prediction of Clinical Deterioration in Hospitalized Adult Patients with Hematologic Malignancies Using a Neural Network Model , 2016, PloS one.

[8]  Majid Sarrafzadeh,et al.  Building Continuous Arterial Blood Pressure Prediction Models Using Recurrent Networks , 2016, 2016 IEEE International Conference on Smart Computing (SMARTCOMP).

[9]  Maysam F. Abbod,et al.  Computational Depth of Anesthesia via Multiple Vital Signs Based on Artificial Neural Networks , 2015, BioMed research international.

[10]  Maysam F. Abbod,et al.  Sample Entropy Analysis of EEG Signals via Artificial Neural Networks to Model Patients' Consciousness Level Based on Anesthesiologists Experience , 2015, BioMed research international.

[11]  M. Jeanne,et al.  Analgesia/nociception index for the assessment of acute postoperative pain. , 2014, British journal of anaesthesia.

[12]  B. Bein,et al.  Influence of nociceptive stimulation on analgesia nociception index (ANI) during propofol-remifentanil anaesthesia. , 2013, British journal of anaesthesia.

[13]  Régis Logier,et al.  Heart rate variability analysis as an index of emotion regulation processes: Interest of the Analgesia Nociception Index (ANI). , 2012, 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[14]  V. Bonhomme,et al.  Effect of a fluid challenge on the Surgical Pleth Index during stable propofol–remifentanil anaesthesia , 2012, Acta anaesthesiologica Scandinavica.

[15]  Régis Logier,et al.  Variations of the analgesia nociception index during general anaesthesia for laparoscopic abdominal surgery , 2012, Journal of Clinical Monitoring and Computing.

[16]  A. Dassonneville,et al.  PhysioDoloris: a monitoring device for Analgesia / Nociception balance evaluation using Heart Rate Variability analysis , 2010, 2010 Annual International Conference of the IEEE Engineering in Medicine and Biology.

[17]  J. De Jonckheere,et al.  Validation of a graphic measurement of heart rate variability to assess analgesia/nociception balance during general anesthesia , 2009, 2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[18]  Geoffrey J. Gordon,et al.  Artificial intelligence in medicine , 1989, Singapore medical journal.

[19]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[20]  Mustafa Gazi,et al.  [Analgesia Nociception Index for perioperative analgesia monitoring in spinal surgery]. , 2017, Revista brasileira de anestesiologia.

[21]  A. Dassonneville,et al.  An efficient algorithm for R-R intervals series filtering , 2004, The 26th Annual International Conference of the IEEE Engineering in Medicine and Biology Society.