Empirical mode decomposition and neural network for the classification of electroretinographic data

Comparison of Machine Learning Approaches to Improve Diagnosis of Optic Neuropathy Using Photopic Negative Response Measured Using a Handheld Device

The photopic negative response of the full-field electroretinogram (ERG) is reduced in optic neuropathies. However, technical requirements for measurement and poor classification performance have limited widespread clinical application. Recent advances in hardware facilitate efficient clinic-based recording of the full-field ERG. Time series classification, a machine learning approach, may improve classification by using the entire ERG waveform as the input. In this study, full-field ERGs were recorded in 217 eyes (109 optic neuropathy and 108 controls) of 155 subjects. User-defined ERG features including photopic negative response were reduced in optic neuropathy eyes (p < 0.0005, generalized estimating equation models accounting for age). However, classification of optic neuropathy based on user-defined features was only fair with receiver operating characteristic area under the curve ranging between 0.62 and 0.68 and F1 score at the optimal cutoff ranging between 0.30 and 0.33. In comparison, machine learning classifiers using a variety of time series analysis approaches had F1 scores of 0.58–0.76 on a test data set. Time series classifications are promising for improving optic neuropathy diagnosis using ERG waveforms. Larger sample sizes will be important to refine the models.

Empirical mode decomposition processing to improve multifocal-visual-evoked-potential signal analysis in multiple sclerosis

Objective

To study the performance of multifocal-visual-evoked-potential (mfVEP) signals filtered using empirical mode decomposition (EMD) in discriminating, based on amplitude, between control and multiple sclerosis (MS) patient groups, and to reduce variability in interocular latency in control subjects.

Methods

MfVEP signals were obtained from controls, clinically definitive MS and MS-risk progression patients (radiologically isolated syndrome (RIS) and clinically isolated syndrome (CIS)). The conventional method of processing mfVEPs consists of using a 1–35 Hz bandpass frequency filter (X_DFT).

The EMD algorithm was used to decompose the X_DFT signals into several intrinsic mode functions (IMFs). This signal processing was assessed by computing the amplitudes and latencies of the X_DFT and IMF signals (X_EMD). The amplitudes from the full visual field and from ring 5 (9.8–15° eccentricity) were studied. The discrimination index was calculated between controls and patients. Interocular latency values were computed from the X_DFT and X_EMD signals in a control database to study variability.

Results

Using the amplitude of the mfVEP signals filtered with EMD (X_EMD) obtains higher discrimination index values than the conventional method when control, MS-risk progression (RIS and CIS) and MS subjects are studied. The lowest variability in interocular latency computations from the control patient database was obtained by comparing the X_EMD signals with the X_DFT signals. Even better results (amplitude discrimination and latency variability) were obtained in ring 5 (9.8–15° eccentricity of the visual field).

Conclusions

Filtering mfVEP signals using the EMD algorithm will result in better identification of subjects at risk of developing MS and better accuracy in latency studies. This could be applied to assess visual cortex activity in MS diagnosis and evolution studies.

Machine Learning Techniques in Clinical Vision Sciences

This review presents and discusses the contribution of machine learning techniques for diagnosis and disease monitoring in the context of clinical vision science. Many ocular diseases leading to blindness can be halted or delayed when detected and treated at its earliest stages. With the recent developments in diagnostic devices, imaging and genomics, new sources of data for early disease detection and patients' management are now available. Machine learning techniques emerged in the biomedical sciences as clinical decision-support techniques to improve sensitivity and specificity of disease detection and monitoring, increasing objectively the clinical decision-making process. This manuscript presents a review in multimodal ocular disease diagnosis and monitoring based on machine learning approaches. In the first section, the technical issues related to the different machine learning approaches will be present. Machine learning techniques are used to automatically recognize complex patterns in a given dataset. These techniques allows creating homogeneous groups (unsupervised learning), or creating a classifier predicting group membership of new cases (supervised learning), when a group label is available for each case. To ensure a good performance of the machine learning techniques in a given dataset, all possible sources of bias should be removed or minimized. For that, the representativeness of the input dataset for the true population should be confirmed, the noise should be removed, the missing data should be treated and the data dimensionally (i.e., the number of parameters/features and the number of cases in the dataset) should be adjusted. The application of machine learning techniques in ocular disease diagnosis and monitoring will be presented and discussed in the second section of this manuscript. To show the clinical benefits of machine learning in clinical vision sciences, several examples will be presented in glaucoma, age-related macular degeneration, and diabetic retinopathy, these ocular pathologies being the major causes of irreversible visual impairment.

Effect of fuzzy partitioning in Crohn's disease classification: a neuro-fuzzy-based approach

Crohn's disease (CD) diagnosis is a tremendously serious health problem due to its ultimately effect on the gastrointestinal tract that leads to the need of complex medical assistance. In this study, the backpropagation neural network fuzzy classifier and a neuro-fuzzy model are combined for diagnosing the CD. Factor analysis is used for data dimension reduction. The effect on the system performance has been investigated when using fuzzy partitioning and dimension reduction. Additionally, further comparison is done between the different levels of the fuzzy partition to reach the optimal performance accuracy level. The performance evaluation of the proposed system is estimated using the classification accuracy and other metrics. The experimental results revealed that the classification with level-8 partitioning provides a classification accuracy of 97.67 %, with a sensitivity and specificity of 96.07 and 100 %, respectively.

Multi-dimensional complete ensemble empirical mode decomposition with adaptive noise applied to laser speckle contrast images

Laser speckle contrast imaging (LSCI) is a noninvasive full-field optical technique which allows analyzing the dynamics of microvascular blood flow. LSCI has attracted attention because it is able to image blood flow in different kinds of tissue with high spatial and temporal resolutions. Additionally, it is simple and necessitates low-cost devices. However, the physiological information that can be extracted directly from the images is not completely determined yet. In this work, a novel multi-dimensional complete ensemble empirical mode decomposition with adaptive noise (MCEEMDAN) is introduced and applied in LSCI data recorded in three physiological conditions (rest, vascular occlusion and post-occlusive reactive hyperaemia). MCEEMDAN relies on the improved complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and our algorithm is specifically designed to analyze multi-dimensional data (such as images). Over the recent multi-dimensional ensemble empirical mode decomposition (MEEMD), MCEEMDAN has the advantage of leading to an exact reconstruction of the original data. The results show that MCEEMDAN leads to intrinsic mode functions and residue that reveal hidden patterns in LSCI data. Moreover, these patterns differ with physiological states. MCEEMDAN appears as a promising way to extract features in LSCI data for an improvement of the image understanding.