Machine Learning Techniques for the Diagnosis of Attention-Deficit/Hyperactivity Disorder from Magnetic Resonance Imaging: A Concise Review.

BACKGROUND: Attention-deficit/hyperactivity disorder (ADHD) is a neuro-developmental disease commonly seen in children and it is diagnosed via extensive interview procedures, behavioral studies, third-party observations, and comprehensive personal history. ADHD causes regional atrophy in brain regions and alters the pattern of functional brain connectivity networks. Automated/computerized methods based on magnetic resonance imaging (MRI) can replace subjective methods for the identification of ADHD. OBJECTIVES: The aim of this study was to analyze various machine-learning algorithms for ADHD by feeding in vital input features extracted from functional brain connectivity and different existing methods and to review factors crucial for the diagnosis of ADHD. METHODS: This paper is a concise review of machine learning methods for the diagnosis of ADHD from MRI. Techniques for feature extraction, dimensionality reduction/feature selection, and classification, employed in the computerized techniques for the diagnosis of ADHD from MRI and the accuracy of classification offered by the individual methods, are focussed on the review. CONCLUSIONS: Machine learning algorithms with features of functional brain connectivity networks as input, with hierarchical sparse feature elimination, exhibits the highest accuracy. Augmentation of the behavioral features does not contribute much to increased accuracy. The level of accuracy offered by the frameworks meant for the computer-aided diagnosis of ADHD, available in the literature, does not justify their feasibility in clinical practice. Computerized methods that exploit highly specific biomarkers of ADHD like brain iron concentration in Globus Pallidus, Putamen, Caudate nucleus, and thalamus as features are not available.

Respiratory Sound Based Classification of Chronic Obstructive Pulmonary Disease: a Risk Stratification Approach in Machine Learning Paradigm.

This article investigates the classification of normal and COPD subjects on the basis of respiratory sound analysis using machine learning techniques. Thirty COPD and 25 healthy subject data are recorded. Total of 39 lung sound features and 3 spirometry features are extracted and evaluated. Various parametric and nonparametric tests are conducted to evaluate the relevance of extracted features. Classifiers such as support vector machine (SVM), k-nearest neighbor (KNN), logistic regression (LR), decision tree and discriminant analysis (DA) are used to categorize normal and COPD breath sounds. Classification based on spirometry parameters as well as respiratory sound parameters are assessed. Maximum classification accuracy of 83.6% is achieved by the SVM classifier while using the most relevant lung sound parameters i.e. median frequency and linear predictive coefficients. Further, SVM classifier and LR classifier achieved classification accuracy of 100% when relevant lung sound parameters, i.e. median frequency and linear predictive coefficient are combined with the spirometry parameters, i.e. forced vital capacity (FVC) and forced expiratory volume in 1 s (FEV1). It is concluded that combining lung sound based features with spirometry data can improve the accuracy of COPD diagnosis and hence the clinician's performance in routine clinical practice. The proposed approach is of great significance in a clinical scenario wherein it can be used to assist clinicians for automated COPD diagnosis. A complete handheld medical system can be developed in the future incorporating lung sounds for COPD diagnosis using machine learning techniques.

Savitzky-Golay Filter for Denoising Lung Sound

ABSTRACT For computerized analysis of respiratory sounds to be effective, the acquired signal must be free from all the interfering elements. Different forms of noise which can degrade the quality of lung sounds are recording artifacts, power line/Radio Frequency (RF) interferences, ambient acoustic interferences, heart sound interference etc. Such interferences adversely affect the diagnostic interpretations. Powerful denoising techniques are necessary to resolve this issue. A denoising scheme for lung sounds, based on Savitzky-Golay (S-G) filter is proposed in this paper. The order and frame length of the SG filter is determined objectively using the Signal to Noise Ratio (SNR) and computational time as objective function. Maximum SNR is observed when the frame length is nearest to the value just higher than the polynomial order. This observation holds good for different levels of simulated addictive Gaussian noise. The polynomial order of 8 and frame size of 9 are found to be promising with SNR of 10.401db at computation time of 2.1ms.

Modified limb lead ECG system effects on electrocardiographic wave amplitudes and frontal plane axis in sinus rhythm subjects.

OBJECTIVE: Modified Limb Lead (MLL) ECG system may be used during rest or exercise ECG, or atrial activity enhancement. Because of modification in the limb electrode placement, changes are likely to happen in ECG wave amplitudes and frontal plane axis, which may alter the clinical limits of normality and ECG diagnostic criteria. The present study investigated the effects of the modified limb electrode position on the electrocardiographic waveforms, ST segment amplitudes (STa) and frontal plane axis. METHODS: The observational study included sixty sinus rhythm subjects of mean age 38.85±8.76 (SD) in the range 25 to 58 years. In addition to 12-lead ECG, MLL ECG was recorded with, the RA electrode placed in the 3rd right intercostal space to the right of the parasternal line, the LA electrode placed in the 5th right intercostal space to the right of the mid-clavicular line and the LL electrode placed in the 5th right intercostal space on the mid-clavicular line. RESULTS: The modification produced profound changes in ECG wave amplitudes and STa amplitudes in frontal plane leads. The QRS and T wave axis shifted on the average by -17o and 41o, respectively, with considerable individual variation, which altered the diagnostic criteria. CONCLUSION: The ECG amplitudes and STa changes produced by the MLL system showed that all remains within the clinical limits, except the R wave amplitude in the modified lead I. It is evident that the MLL system produced deviations in frontal plane QRS axis which altered the diagnostic interpretation.

The effect of foot arch on plantar pressure distribution during standing.

The aim of this study was to explore how foot type affects plantar pressure distribution during standing. In this study, 32 healthy subjects voluntarily participated and the subject feet were classified as: normal feet (n = 23), flat feet (n = 14) and high arch feet (n = 27) according to arch index (AI) values obtained from foot pressure intensity image analysis. Foot pressure intensity images were acquired by a pedopowergraph system to obtain a foot pressure distribution parameter-power ratio (PR) during standing in eight different regions of the foot. Contact area and mean PR were analysed in hind foot, mid-foot and fore foot regions. One-way analysis of variance was used to determine statistical differences between groups. The contact area and mean PR value beneath the mid-foot was significantly increased in the low arch foot when compared to the normal arch foot and high arch foot (p < 0.001) in both feet. However, subjects with low-arch feet had significantly higher body mass index (BMI) compared to subjects with high-arch feet (p < 0.05) and subjects with normal arch feet (p < 0.05) in both feet. In addition, subjects with low-arch feet had significant differences in arch index (AI) value as compared to subjects with high-arch feet (p < 0.001) and subjects with normal arch feet (p < 0.05) in both feet. Mean mid-foot PR value were positively (r = 0.54) correlated with increased arch index (AI) value. A significant (p < 0.05) change was obtained in PR value beneath the mid-foot of low arch feet when compared with other groups in both feet. The findings suggest that there is an increased mid-foot PR value in the low arch foot as compared to the normal arch foot and high arch foot during standing. Therefore, individuals with low arch feet could be at high risk for mid-foot collapse and Charcot foot problems, indicating that foot type should be assessed when determining an individual's risk for foot injury.

Prevalence of standing plantar pressure distribution variation in north Asian Indian patients with diabetes mellitus: a study to understand ulcer formation.

Diabetes Mellitus is a disorder of metabolism. Foot problems are common in diabetes and altered plantar pressures distribution may lead to ulceration in people with Diabetes Mellitus. Therefore the aim of this study was to investigate standing plantar pressure distribution variations in north Asian Indian diabetes mellitus subjects and its association with duration of diabetes. Thirty three subjects with age range from 40 to 75 years are recruited from AIIMS Endocrinology & metabolism lab Delhi, India and divided into three groups: 11 control subjects (non-diabetic), 11 diabetic subjects without neuropathy (DNN) and II diabetic subjects with neuropathy (DN). Neuropathy status was assessed by measuring loss of protective sensation to 10 gm Semen's Weinstein monofilament. Plantar pressure distributions parameter-Power ratio (PR) was measured during barefoot standing using portable PedoPowerGraph and results are analyzed using one way analysis of variance to detect significant difference between the groups. We found significant (p < 0.05; p < 0.01) difference in PR value between DN and CG groups in fore foot and hind foot but no significant (p > 0.05) difference in PR value was found between DNN and CG groups in the foot. As compared to DNN, DN group have maximum PR variations in the fore foot. Plantar pressure distribution parameter-PR was higher with longer duration of diabetes among type 2 diabetes subjects. In this study we conclude that plantar pressure distribution parameter-PR was able to distinguish the DN groups from the CG group in hind and fore foot during standing. Increased forefoot PR value is prevalent in the diabetic neuropathic subjects and may be responsible for the occurrence of foot sole ulcers but additional prospective studies are needed. In the future we will investigate the plantar pressure distribution parameter-PR variations in diabetes with obese and osteoarthritis subject.

Foot pressure distribution variation in pre-obese and non-obese adult subject while standing.

BACKGROUND: To investigate foot pressure distribution parameter-power ratio (PR) difference between pre-obese and non-obese adults subjects during standing and show the correlation between body mass index (BMI) and PR value. METHODS: We examined 22 healthy adult subjects aged from 20 to 45 years were classified into two groups according to their BMI values, as 11 non-obese and 11 pre-obese subjects. Foot pressure distribution image during standing was obtained using PedoPowerGraph system. Pedopowergraphic parameters such as percentage medial impulse, forefoot to hind foot pressure distribution ratio and PR were evaluated and compared between the groups. Correlation between BMI value and PR value was assessed. RESULTS: Our result shows significant change in contact area between the groups in mid foot regions. Also we found significant differences in mid foot PR values (p<0.05) between the groups, but no significant differences in hind foot and forefoot PR values. In addition BMI value was found to have positive correlation with right and left mid foot PR value (r=0.60 & 0.61) for all the subjects. CONCLUSION: This study provides for the first time new insights into foot pressure distribution difference in mid foot among pre-obese subjects as compared to non-obese adult subject while standing. Hence knowledge of high mid foot PR value among pre-obese subjects can provide suitable guidelines for designing orthotic devices.

Association of limited joint mobility and increased plantar hardness in diabetic foot ulceration in north Asian Indian: a preliminary study.

The aim of this article is to investigate the association of limited joint mobility and foot sole hardness in north Asian Indian type 2 diabetic patients. Limited joint mobility and hardness of the foot sole were measured for 39 subjects attending the AIIMS Endocrinology & Metabolism Clinic. The total subject divided into three groups: 13 control subjects (nondiabetic), 13 diabetic patients without neuropathy and 13 diabetic neuropathy patients. Neuropathy status was assessed using 10 gm Semen's Weinstein monofilament. Joint mobility parameters, such as ankle dorsiflexion/plantar flexion and metatarsophalangeal-1 dorsiflexion/plantar flexion, are measured using a goniometer. Foot sole hardness was measured using a durometer or shore meter. We found that diabetic patients with a neuropathic foot had significantly reduced joint mobility and increased foot sole hardness, placing them at risk for subsequent ulceration. Metatarsophalangeal-1 dorsiflexion/plantar flexion of both feet of diabetic patients had significant correlation (at p < 0.05, p < 0.001, p < 0.001 level) over age and body mass index. Also ankle plantar flexion/dorsiflexion and metatarsophalangeal-1 dorsiflexion/plantar flexion has a significant correlations (at p < 0.01, p < 0.05, p < 0.001, p < 0.001 level) with foot sole hardness in both feet of diabetic neuropathy subjects. Also linear regression analysis showed that duration of diabetes was significantly associated with the joint mobility parameters. In this study we conclude that joint mobility had reduced further if neuropathy and increased foot sole hardness coexisted owing to high plantar pressures. Hence, both limited joint mobility and increased foot sole hardness appears to be important determinants of foot sole ulceration in diabetic neuropathic subject.

The effect of aging on the hardness of foot sole skin: a preliminary study.

BACKGROUND: Foot problems are common in older people and altered biomechanical parameters under the foot sole has been proposed as a key risk factor for foot lesions. Therefore the aim of this study was to investigate the age-related differences in the hardness of foot sole skin. METHODS: Twenty-six healthy volunteers without foot problems, aged from 26 to 65 years, were examined using shore meter. The hardness of the foot sole under the big toe (area 8), 1st metatarsal head (area 5), 3rd metatarsal head (area 6), 5th metatarsal head (area 7), mid foot (area 3, 4) and hind foot (area 1, 2) were measured. The correlation between age and hardness of foot sole was examined and comparisons were made between two age groups. RESULTS: From the result we observe statistical significant (p<0.05; p<0.01; p<0.005) differences in hardness between age groups in hind foot, metatarsal heads (1st, 3rd & 5th) and big toe. Strong positive correlations between age and hardness of the foot sole were found at the big toe (r=0.57; p<0.005), 1st metatarsal head (r=0.567; p<0.00001), 3rd metatarsal head (r=0.565; p<0.00001), 5th metatarsal head (r=0.55; p<0.00001), and heel (r=0.59; p<0.0001). CONCLUSION: The loss of compliance in the foot sole may be one of the factors responsible for the higher incidence of foot problems in aged people. Routine foot examination and appropriate therapeutic intervention including the use of foot orthoses and optimal hardness of foot wear insole may help to prevent the serious foot injuries.

Preliminary investigation of foot pressure distribution variation in men and women adults while standing.

BACKGROUND: Women and men are anatomically and physiologically different in a number of ways. They differ in both shape and size. These differences could potentially mean foot pressure distribution variation in men and women. The purpose of this study was to analyze standing foot pressure image to obtain the foot pressure distribution parameter - power ratio variation between men and women using image processing in frequency domain. METHODS: We examined 28 healthy adult subjects (14 men and 14 women) aged between 20 and 45 years was recruited for our study. Foot pressure distribution patterns while standing are obtained by using a PedoPowerGraph plantar pressure measurement system for foot image formation, a digital camera for image capturing, a TV tuner PC-add on card, a WinDvr software for still capture and Matlab software with dedicated image processing algorithms have been developed. Various PedoPowerGraphic parameters such as percentage medial impulse (PMI), fore foot to hind foot pressure distribution ratio (F/H), big toe to fore foot pressure distribution ratio (B/F) and power ratio (PR) were evaluated. RESULTS: In men, contact area was significantly larger in all regions of the foot compared with women. There were significant differences in plantar pressure distribution but there was no significant difference in F/H and B/F ratio. Mean PR value was significantly greater in men than women under the hind foot and fore foot. PMI value was greater in women than men. As compared to men, women have maximum PR variations in the mid foot. Hence there is significant difference at level p<0.05 in medial mid foot and mid foot PR of women as compared to men. CONCLUSION: There was variation in plantar pressure distribution because the contact area of the men foot was larger than that of women foot. Hence knowledge of pressure distributions variation of both feet can provide suitable guidelines to biomedical engineers and doctor for designing orthotic devices for reliving the area of excessively high pressure.