Shubhajit Roy Chowdhury

Accuracy Enhancement in a Fuzzy Expert Decision Making System Through Appropriate Determination of Membership Functions and Its Application in a Medical Diagnostic Decision Making System

The paper attempts to improve the accuracy of a fuzzy expert decision making system by tuning the parameters of type-2 sigmoid membership functions of fuzzy input variables and hence determining the most appropriate type-1 membership function. The current work mathematically models the variability of human decision making process using type-2 fuzzy sets. Moreover, an index of accuracy of a fuzzy expert system has been proposed and determined analytically. It has also been ascertained that there exists only one rule in the rule base whose associated mapping for the ith linguistic variable maps to the same value as the maximum value of the membership function for the ith linguistic variable. The improvement in decision making accuracy was successfully verified in a medical diagnostic decision making system for renal diagnostic applications. Based on the accuracy estimations applied over a set of pathophysiological parameters, viz. body mass index, glucose, urea, creatinine, systolic and diastolic blood pressure, appropriate type-1 fuzzy sets of these parameters have been determined assuming normal distribution of type-1 membership function values in type-2 fuzzy sets. The type-1 fuzzy sets so determined have been used to develop an FPGA based smart processor. Using the processor, renal diagnosis of patients has been performed with an accuracy of 98.75%.

FPGA realization of a smart processing system for clinical diagnostic applications using pipelined datapath architectures

The paper describes the implementation of a novel smart data processing hardware system having medical diagnostic capabilities. The system is implemented in a FPGA chip for fast design cycle. The system employs a smart agent that can predict the future pathophysiological state of a patient using the past pathophysiological data and can thus predict the approaching critical condition of the patient before criticality actually occurs. This feature has been realized using fuzzy logic. In order to speed up the computation process, pipelined data processing architectures have been employed leading to a speed up of approximately three times. The whole system is realized on Altera Cyclone EP1K6Q240C8 FPGA chip requiring 5308 logic blocks. The system has been designed to be inexpensive, portable and user friendly for rural applications in developing countries.

Development of Point of Care Testing Device for Neurovascular Coupling From Simultaneous Recording of EEG and NIRS During Anodal Transcranial Direct Current Stimulation

This paper presents a point of care testing device for neurovascular coupling (NVC) from simultaneous recording of electroencephalogram (EEG) and near infrared spectroscopy (NIRS) during anodal transcranial direct current stimulation (tDCS). Here, anodal tDCS modulated cortical neural activity leading to hemodynamic response can be used to identify the impaired cerebral microvessels functionality. The impairments in the cerebral microvessels functionality may lead to impairments in the cerebrovascular reactivity (CVR), where severely reduced CVR predicts the chances of transient ischemic attack and ipsilateral stroke. The neural and hemodynamic responses to anodal tDCS were studied through joint imaging with EEG and NIRS, where NIRS provided optical measurement of changes in tissue oxy-(HbO2) and deoxy-(Hb) hemoglobin concentration and EEG captured alterations in the underlying neuronal current generators. Then, a cross-correlation method for the assessment of NVC underlying the site of anodal tDCS is presented. The feasibility studies on healthy subjects and stroke survivors showed detectable changes in the EEG and the NIRS responses to a 0.526 A/m2 of anodal tDCS. The NIRS system was bench tested on 15 healthy subjects that showed a statistically significant (p <; 0.01) difference in the signal-to-noise ratio (SNR) between the ONand OFF-states of anodal tDCS where the mean SNR of the NIRS device was found to be 42.33 ± 1.33 dB in the ON-state and 40.67±1.23 dB in the OFF-state. Moreover, the clinical study conducted on 14 stroke survivors revealed that the lesioned hemisphere with impaired circulation showed significantly (p <; 0.01) less change in HbO2 than the nonlesioned side in response to anodal tDCS. The EEG study on healthy subjects showed a statistically significant (p <; 0.05) decrease around individual alpha frequency in the alpha band (8-13 Hz) following anodal tDCS. Moreover, the joint EEG-NIRS imaging on 4 stroke survivors showed an immediate increase in the theta band (4-8 Hz) EEG activity after the start of anodal tDCS at the nonlesioned hemisphere. Furthermore, cross-correlation function revealed a significant (95% confidence interval) negative cross correlation only at the nonlesioned hemisphere during anodal tDCS, where the log-transformed mean-power of EEG within 0.5-11.25 Hz lagged HbO2 response in one of the stroke survivors with white matter lesions. Therefore, it was concluded that the anodal tDCS can perturb the local neural and the vascular activity (via NVC) which can be used for assessing regional NVC functionality where confirmatory clinical studies are required.

A High-Performance FPGA-Based Fuzzy Processor Architecture for Medical Diagnosis

An auto-decision-making system for medical diagnosis could help make up for the lack of physicians in rural areas of many third-world countries. This high-performance, low-power, pipelined parallel fuzzy processor based on a dedicated single-chip architecture performs high-speed fuzzy inferences with processing speed up to 5.0 Mflips at a clock frequency of 40 MHz using 256 rules having one consequent each, 16 input variables, and 16-bit resolution. The processor operates in real time producing results within an interval of 1.92mus. The processor implemented on board consumes as low as 70 milliwatt power. The processor performs medical diagnosis with 97.5 percent accuracy.

Medical Diagnosis Using Adaptive Perceptive Particle Swarm Optimization and Its Hardware Realization using Field Programmable Gate Array

The paper proposes to develop a field programmable gate array (FPGA) based low cost, low power and high speed novel diagnostic system that can detect in absence of the physician the approaching critical condition of a patient at an early stage and is thus suitable for diagnosis of patients in the rural areas of developing countries where availability of physicians and availability of power is really scarce. The diagnostic system could be installed in health care centres of rural areas where patients can register themselves for periodic diagnoses and thereby detect potential health hazards at an early stage. Multiple pathophysiological parameters with different weights are involved in diagnosing a particular disease. A novel variation of particle swarm optimization called as adaptive perceptive particle swarm optimization has been proposed to determine the optimal weights of these pathophysiological parameters for a more accurate diagnosis. The FPGA based smart system has been applied for early detection of renal criticality of patients. For renal diagnosis, body mass index, glucose, urea, creatinine, systolic and diastolic blood pressures have been considered as pathophysiological parameters. The detection of approaching critical condition of a patient by the instrument has also been validated with the standard Cockford Gault Equation to verify whether the patient is really approaching a critical condition or not. Using Bayesian analysis on the population of 80 patients under study an accuracy of up to 97.5% in renal diagnosis has been obtained.

Design, Simulation and Testing of a High Speed Low Power 15-4 Compressor for High Speed Multiplication Applications

The paper presents a novel high speed and low power 15-4 Compressor for high speed and low power multiplication applications. The proposed compressor uses bit sliced adder architecture to exploit the parallelism in the computation of sum of 15 input bits by five full adders. The newly proposed compressor is also centered around the design of a novel 5-3 compressor that attempts to minimize the stage delays of a conventional 5-3 compressor that is designed using single bit full adder and half adder architectures. The proposed 15-4 compressor uses the minimum number of hardware resources so far as the logic level architecture of the design is concerned. The proposed compressor is tested using 14 transistor and 10 transistor adder designs reported in literature. The power delay product of the proposed compressor is found to be equal to as low as 0.98fJ using 10T adder design and 0.46fJ using 14T adder design. The whole simulation has been carried out using TSPICE using 0.35 mum technology.

Development of a FPGA based fuzzy neural network system for early diagnosis of critical health condition of a patient

The paper describes the design and training of a fuzzy neural network used for early diagnosis of a patient through an FPGA based implementation of a smart instrument. The system employs a fuzzy interface cascaded with a feed-forward neural network. In order to obtain an optimum decision regarding the future pathophysiological state of a patient, the optimal weights of the synapses between the neurons have been determined by using inverse delayed function model of neurons. The neurons that are considered in the proposed network are devoid of self connections instead of commonly used self connected neurons. The current work also find out the optimal number of neurons in the hidden layer for accurate diagnosis as against the available number of CLB in the FPGA. The system has been trained and tested with renal data of patients taken at 10 days interval of time. Applying the methodology, the chance of attainment of critical renal condition of a patient has been predicted with an accuracy of 95.2%, 30 days ahead of actually attaining the critical condition. The system has also been tested for pathophysiological state prediction of patients at multiple time steps ahead and the prediction at the next instant of time stands out to be the most accurate.

A phenomological model for capturing cerebrovascular reactivity to anodal transcranial direct current stimulation

The paper presents a phenomological model to capture cerebrovascular reactivity (CVR) that represented the capacity of blood vessels to dilate during anodal transcranial direct current stimulation (tDCS). Anodal tDCS modulated cortical neural activity leading to CVR, where it can identify impaired cerebral microvessels functionality leading to impairments of cerebral blood flow that may cause impairments in the cerebral functions. In this study, CVR was probed with near infra-red spectroscopy (NIRS) where NIRS recorded changes in oxy-haemoglobin and deoxy-haemoglobin concentrations during anodal tDCS-induced activation of the cortical region located under the electrode and in-between the light sources and detectors. The regional CVR during anodal tDCS was captured by adapting an arteriolar compliance model of the cerebral blood flow response to neural stimuli, where a fourth-order discrete-time model represented the haemodynamic response to anodal tDCS. A case study showed detectable CVR response (0-60sec) to a 0.526A/m2 square-pulse (0-30sec) of anodal tDCS where these alterations in the vascular system may result in secondary changes in the cortical excitability. This needs to be carefully studied in the future with multi-modal imaging in a larger patient group, for example, in cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy.

Corticospinal excitability changes to anodal tDCS elucidated with NIRS-EEG joint-imaging: An ischemic stroke study.

Transcranial direct current stimulation (tDCS) has been shown to modulate corticospinal excitability. We used near-infrared spectroscopy (NIRS) - electroencephalography (EEG) joint-imaging during and after anodal tDCS to measure changes in mean cerebral haemoglobin oxygen saturation (rSO2) along with changes in the log-transformed mean-power of EEG within 0.5 Hz - 11.25 Hz. In two separate studies, we investigated local post-tDCS alterations from baseline at the site of anodal tDCS using NIRS-EEG/tDCS joint-imaging as well as local post-tDCS alterations in motor evoked potentials (MEP)-measure of corticospinal excitability. In the first study, we found that post-tDCS changes in the mean rSO2 from baseline mostly correlated with the corresponding post-tDCS change in log-transformed mean-power of EEG within 0.5 Hz - 11.25 Hz. Moreover, a decrease in log-transformed mean-power of EEG within 0.5 Hz - 11.25 Hz corresponded with an increase in the MEP-measure of corticospinal excitability - found in the second study. Therefore, we propose to combine NIRS-EEG/tDCS joint-imaging with corticospinal excitability investigation in a single study to confirm these finding. Furthermore, we postulate that the innovative technologies for portable NIRS-EEG neuroimaging may be leveraged to objectively quantify the progress (e.g., corticospinal excitability alterations) and dose tDCS intervention as an adjuvant treatment during neurorehabilitation.

Low Power Motion Estimation Based on Probabilistic Computing

As CMOS technology driven by Moores law has approached device sizes in the range of 5-20 nm, noise immunity of such future technology nodes is predicted to decrease considerably, eventually affecting the reliability of computations through them. A shift in the design paradigm is expected from 100% accurate computations to probabilistic computing with accuracy dependent on the target application or circuit specifications. One model developed for CMOS technology that emulates the erroneous behavior predicted is termed probabilistic CMOS (PCMOS). In this paper, we propose a PCMOS-based architecture implementation for traditional motion estimation algorithms and show that up to 57% energy savings are possible for different existing motion estimation algorithms. Furthermore, algorithmic modifications are proposed that can enhance the energy savings to 70% with a PCMOS architectural implementation. About 1.8-5 dB improvement in peak signal-to-noise ratio under energy savings of 57% to 70% for two different motion estimation algorithms is shown, establishing the resilience of the proposed algorithm to probabilistic computing over the comparable conventional algorithm.