Artificial neural network to assist psychiatric diagnosis

Proof of concept: Predicting distress in cancer patients using back propagation neural network (BPNN)

Background

Research findings suggest that a significant proportion of individuals diagnosed with cancer, ranging from 25% to 60%, experience distress and require access to psycho-oncological services. Until now, only contemporary approaches, such as logistic regression, have been used to determine predictors of distress in oncological patients. To improve individual prediction accuracy, novel approaches are required. We aimed to establish a prediction model for distress in cancer patients based on a back propagation neural network (BPNN).

Methods

Retrospective data was gathered from a cohort of 3063 oncological patients who received diagnoses and treatment spanning the years 2011-2019. The distress thermometer (DT) has been used as screening instrument. Potential predictors of distress were identified using logistic regression. Subsequently, a prediction model for distress was developed using BPNN.

Results

Logistic regression identified 13 significant independent variables as predictors of distress, including emotional, physical and practical problems. Through repetitive data simulation processes, it was determined that a 3-layer BPNN with 8 neurons in the hidden layer demonstrates the highest level of accuracy as a prediction model. This model exhibits a sensitivity of 79.0%, specificity of 71.8%, positive predictive value of 78.9%, negative predictive value of 71.9%, and an overall coincidence rate of 75.9%.

Conclusion

The final BPNN model serves as a compelling proof of concept for leveraging artificial intelligence in predicting distress and its associated risk factors in cancer patients. The final model exhibits a remarkable level of discrimination and feasibility, underscoring its potential for identifying patients vulnerable to distress.

Reading Wishes from the Lips: Cancer Patients' Need for Psycho-Oncological Support during Inpatient and Outpatient Treatment

Background: Psycho-oncological support (PO) is an effective measure to reduce distress and improve the quality of life in patients with cancer. Currently, there are only a few studies investigating the (expressed) wish for PO. The aim of this study was to evaluate the number of patients who request PO and to identify predictors for the wish for PO. Methods: Data from 3063 cancer patients who had been diagnosed and treated at a Comprehensive Cancer Center between 2011 and 2019 were analyzed retrospectively. Potential predictors for the wish for PO were identified using logistic regression. As a novelty, a Back Propagation Neural Network (BPNN) was applied to establish a prediction model for the wish for PO. Results: In total, 1752 patients (57.19%) had a distress score above the cut-off and 14.59% expressed the wish for PO. Patients’ requests for pastoral care (OR = 13.1) and social services support (OR = 5.4) were the strongest predictors of the wish for PO. Patients of the female sex or who had a current psychiatric diagnosis, opioid treatment and malignant neoplasms of the skin and the hematopoietic system also predicted the wish for PO, while malignant neoplasms of digestive organs and older age negatively predicted the wish for PO. These nine significant predictors were used as input variables for the BPNN model. BPNN computations indicated that a three-layer network with eight neurons in the hidden layer is the most precise prediction model. Discussion: Our results suggest that the identification of predictors for the wish for PO might foster PO referrals and help cancer patients reduce barriers to expressing their wish for PO. Furthermore, the final BPNN prediction model demonstrates a high level of discrimination and might be easily implemented in the hospital information system.

Study on quantitative diagnosis model of TCM syndromes of post-stroke depression based on combination of disease and syndrome

BACKGROUND

Post-stroke depression (PSD) is one of the most common stroke complications with high morbidity. Researchers have done much clinical research on Traditional Chinese Medicine (TCM) treatment, but very little research on diagnosis. Based on the thought of combination of disease and syndrome, this study will establish a unified and objective quantitative diagnosis model of TCM syndromes of PSD, so as to improve the clinical diagnosis and treatment of PSD.

OBJECTIVE

First: To establish a unified and objective quantitative diagnosis model of TCM syndromes in PSD under different disease courses, and identify the corresponding main, secondary, and concurrent symptoms, which are based on the weighting factor of each TCM symptom. Second: To find out the relationship between different stages of PSD and TCM syndromes. Clarify the main syndrome types of PSD under different stages of disease. Reveal the evolution and progression mechanism of TCM syndromes of PSD.

METHODS AND ANALYSIS

This is a retrospective study of PSD TCM diagnosis. Three hundred patients who were hospitalized in the First Teaching Hospital of Tianjin University of TCM with complete cases from January 2014 to January 2019 are planned to be recruited. The study will mainly collect the diagnostic information from the cases, find the related indicators of TCM and Western medicine in PSD, and clarify the relationship between different disease stages and TCM syndromes. Finally, the PSD TCM syndrome quantitative diagnosis model will be established based on the operation principle of Back Propagation (BP) artificial neural network.

CONCLUSION

To collect sufficient medical records and establish models to speed up the process of TCM diagnosis.

Neural network approaches to grade adult depression

Depression is a common but worrying psychological disorder that adversely affects one's quality of life. It is more ominous to note that its incidence is increasing. On the other hand, screening and grading of depression is still a manual and time consuming process that might be biased. In addition, grades of depression are often determined in continuous ranges, e.g., 'mild to moderate' and 'moderate to severe' instead of making them more discrete as 'mild', 'moderate', and 'severe'. Grading as a continuous range is confusing to the doctors and thus affecting the management plan at large. Given this practical issue, the present paper attempts to differentiate depression grades more accurately using two neural net learning approaches-'supervised', i.e., classification with Back propagation neural network (BPNN) and Adaptive Network-based Fuzzy Inference System (ANFIS) classifiers, and 'unsupervised', i.e., 'clustering' technique with Self-organizing map (SOM), built in MATLAB 7. The reason for using the supervised and unsupervised learning approaches is that, supervised learning depends exclusively on domain knowledge. Supervision may induce biasness and subjectivities related to the decision-making. Finally, the performance of BPNN and ANFIS are compared and discussed. It was observed that ANFIS, being a hybrid system, performed much better compared to the BPNN classifier. On the other hand, SOM-based clustering technique is also found useful in constructing three distinct clusters. It also assists visualizing the multidimensional data clusters into 2-D.

Designing a decision support system for distinguishing ADHD from similar children behavioral disorders

In this study, a decision support system was designed to distinguish children with ADHD from other similar children behavioral disorders such as depression, anxiety, comorbid depression and anxiety and conduct disorder based on the signs and symptoms. Accuracy of classifying with Radial basis function and multilayer neural networks were compared. Finally, the average accuracy of the networks in classification reached to 95.50% and 96.62% by multilayer and radial basis function networks respectively. Our results indicate that a decision support system, especially RBF, may be a good preliminary assistant for psychiatrists in diagnosing high risk behavioral disorders of children.

A principal component network analysis of prefrontal-limbic functional magnetic resonance imaging time series in schizophrenia patients and healthy controls

We investigated neural regulation of emotional arousal. We hypothesized that the interactions between the components of the prefrontal-limbic system determine the global trajectories of the individual's brain activation, with the strengths and modulations of these interactions being potentially key components underlying the differences between healthy individuals and those with schizophrenia. Using affect-valent facial stimuli presented to 11 medicated schizophrenia patients and 65 healthy controls, we activated neural regions associated with the emotional arousal response during functional magnetic resonance imaging (fMRI). Performing first a random effects analysis of the fMRI data to identify activated regions, we obtained 352 data-point time series for six brain regions: bilateral amygdala, hippocampus and two prefrontal regions (Brodmann Areas 9 and 45). Since standard statistical methods are not designed to capture system features and evolution, we used principal component analyses on two types of pre-processed data: contrasts and group averages. We captured an important characteristic of the evolution of our six-dimensional brain network: all subject trajectories are almost embedded in a two-dimensional plane. Moreover, the direction of the largest principal component was a significant differentiator between the control and patient populations: the left and right amygdala coefficients were substantially higher in the case of patients, and the coefficients of Brodmann Area 9 were, to a lesser extent, higher in controls. These results are evidence that modulations between the regions of interest are the important determinant factors for the system's dynamical behavior. We place our results within the context of other principal component analyses used in neuroimaging, as well as of our existing theoretical model of prefrontal-limbic dysregulation.

Predicting osteoarthritic knee rehabilitation outcome by using a prediction model developed by data mining techniques

Artificial neural networks (ANN) have been applied to assist in clinical decision-making and prediction. While we consider possible effective treatments for patients with osteoarthritic knee such as Transcutaneous Electrical Nerve Stimulation (TENS), exercise, and TENS with exercise respectively, we have to select a treatment protocol for patients such that they would gain the best improvements according to their clinical conditions. To facilitate this functionality with the existing patient assessment, we hope to apply the ANN programming techniques to develop a computerized prediction system. A preliminary validation was performed to test the validity of the newly developed prediction protocol on knee rehabilitation. We input the key clinical attributes of 62 patients who have undergone the three above-mentioned knee treatments to the protocol. The expected pain improvement of each patient as predicted by the protocol was obtained. Spearman rank-order correlation was used to identify whether there was a significant correlation between the rankings of the observed and expected pain improvement. We found that the Spearman's rho was 0.424, which is statistically significant at p < 0.001. From this preliminary analysis, we are confident that this newly developed prediction protocol will be useful when deciding which treatment regime best suits a patient.

Neural network based automatic diagnosis of children with brain dysfunction

This paper proposes the use of multilayer perceptron for brain dysfunction diagnosis. The performance of MLP was better than that of Discriminant Analysis and Decision Tree classifiers, with an 85% accuracy rate in an experimental test involving 332 subjects. In addition, the neural network employing Bayesian learning was able to identify the most important input variable. These two results demonstrate that the neural network can be effectively used in the diagnosis of children with brain dysfunction.

Determining the number of state psychiatric hospital beds by measuring quality of care with artificial neural networks

This study uses a new paradigm to calculate the minimum and the optimum number of involuntary psychiatric beds at a state hospital in Maine with 5538 admissions over a 7-year period. The method measures quality of care (Q) based upon the accuracy of prediction of length-of-stay for the hospital, and of community length-of-stay for the community, each corrected for the severity of illness of the average patient. When Q in the hospital equals Q in the community, there is no net movement of patients from one phase of care to the other, analogous to a zero electromotive force, and the census at that point is the minimum number of beds (22 beds/100,000 population). When patients in the community were least ill, relative to the hospital then hospital bed census is at its optimum (31 beds/100,000) given current resources and technology. In studying specific diagnosis groups with the same methodology the authors found that patients with schizophrenia having the benefit of clozapine for most of the study period had a Q averaged over 7 years that was nearly equal in both hospital and community settings. This explains the perception that tertiary psychiatric hospitals comprised mostly of patients with schizophrenia can downsize significantly. However, affective disorders and "borderline" personality disorders clearly benefit from structured hospital care with specialized experienced staff. We make arguments for the role of the state hospital as a homeostat for the mental health care delivery system.