Md. Abdul Awal

EEG background features that predict outcome in term neonates with hypoxic ischaemic encephalopathy: A structured review

OBJECTIVES Hypoxic ischaemic encephalopathy is a significant cause of mortality and morbidity in the term infant. Electroencephalography (EEG) is a useful tool in the assessment of newborns with HIE. This systematic review of published literature identifies those background features of EEG in term neonates with HIE that best predict neurodevelopmental outcome. METHODS A literature search was conducted using the PubMed, EMBASE and CINAHL databases from January 1960 to April 2014. Studies included in the review described recorded EEG background features, neurodevelopmental outcomes at a minimum age of 12 months and were published in English. Pooled sensitivities and specificities of EEG background features were calculated and meta-analyses were performed for each background feature. RESULTS Of the 860 articles generated by the initial search strategy, 52 studies were identified as potentially relevant. Twenty-one studies were excluded as they did not distinguish between different abnormal background features, leaving 31 studies from which data were extracted for the meta-analysis. The most promising neonatal EEG features are: burst suppression (sensitivity 0.87 [95% CI (0.78-0.92)]; specificity 0.82 [95% CI (0.72-0.88)]), low voltage (sensitivity 0.92 [95% CI (0.72-0.97)]; specificity 0.99 [95% CI (0.88-1.0)]), and flat trace (sensitivity 0.78 [95% CI (0.58-0.91)]; specificity 0.99 [95% CI (0.88-1.0)]). CONCLUSION Burst suppression, low voltage and flat trace in the EEG of term neonates with HIE most accurately predict long term neurodevelopmental outcome. SIGNIFICANCE This structured review and meta-analysis provides quality evidence of the background EEG features that best predict neurodevelopmental outcome.

A robust high-resolution time–frequency representation based on the local optimization of the short-time fractional Fourier transform

The Locally Optimized Spectrogram (LOS) defines a novel method for obtaining a high-resolution time-frequency (t, f) representation based on the short-time fractional Fourier transform (STFrFT). The key novelty of the LOS is that it automatically determines the locally optimal window parameters and fractional order (angle) for all signal components, leading to a high-resolution and cross-terms free time frequency representation. This method is suitable for multicomponent and non-stationary signals without a priori signal information. Simulated signals, real biomedical applications, and various measures are used to validate the improved performance of the LOS and compare it with other state-of-the-art methods. The robustness of the LOS is also demonstrated under different signal-to-noise ratio (SNR) conditions. Finally, the relationship between the LOS and other time-frequency distributions (TFDs) is depicted and a recursive formulation is presented and shows the trade-off between the cross-terms suppression and auto-terms resolution

Detection of neonatal EEG burst-suppression using a time-frequency approach

In newborn EEG, the presence of burst suppression carries with it a high probability of poor neurodevelopmental outcome. This paper presents a novel method to detect neonatal bust suppression from multichannel EEG using a time-frequency (T-F) based approach. In this approach, features are extracted from T-F representations of EEG signals obtained using quadratic time-frequency distributions (QTFDs). Such features take into account the non-stationarity of EEG signals and are shown to be able to discriminate between burst and suppression patterns. The features are based on the energy concentration of the signals in the T-F domain, instantaneous frequency of the signals, and Renyi entropy and singular-value decomposition (SVD) of the TFDs of EEG. For each feature, the receiver operating characteristic (ROC) is found and the area under the ROC curve (AUC) is calculated as the performance criterion. Experimental results using EEG signals with burst suppression acquired from 3 term neonates show that the features extracted from the singular values of TFDs and energy concentration outperform others. Amongst different QTFDs, features extracted from the optimized extended modified B distribution exhibit the best performance. Also, a classifier which uses these features achieves a total accuracy of 99.6%.

Clench force estimation by surface electromyography for neural prosthesis hand

Clench force estimator is highly desirable in the field of prosthesis hand. It is one of the most used postures among five types of postures. In this paper, we propose to estimate the clench force using two types of Surface Electromyography (SEMG). The SMEG consists of rectified SEMG and integrated SEMG. A two layered artificial neural network (ANN) is used as an estimator to map the SEMG for estimating force. For weight adjustment of the estimator Levenberg-Marquardt (L-M) back propagation algorithm is used. The proposed network is trained and tested using SEMG recorded from five subjects. The estimation result clearly show that integrated SEMG performed 3.53 times better than rectified SEMG in the case of cross correlation coefficient and hence integrated SEMG is recommended for clench force estimation.

A Method for Designing Emg Integrator using an FPGA

Electromyography is a method for recording electrical activities of the muscle for different clinical and nonclinical tasks. For extracting more information, integrated electromyography is commonly used than the raw electromyography. This paper presents the design and implementation of integrated electromyography both in software and hardware. Software was implemented in Matlab due to easier implementation whereas hardware was designed on Field Programmable Gate Array (FPGA) due to low cost and flexibility. It can be seen that, the integrator works like a moving average window filter and a hundred-point window size is chosen in the integrator design. To verify the method, real surface electromyography data was collected and used. The mean error between software (Matlab) and hardware is 5.8288e-08 and the correlation coefficient is 1.

Design of sEMG-based clench force estimator in FPGA using artificial neural networks

Hands are the main environmental manipulator for the human being. After losing a hand, the only alternative for the victim is to use a prosthesis. Despite the progress of science, the modern prosthesis has the same age-old problem of accurate force estimation. Among different kinds of force, clench force is the most important one. Because of this importance, this paper presents a hardware system that has been designed and implemented to estimate the desired clench force using surface Electromyography signals recorded from lower-arm muscles. The implementation includes a two-layer artificial neural network with a surface electromyography integrator. The neural network was trained with the Levenberg–Marquardt back propagation algorithm and was implemented in a field programmable gate array using an off-chip training method. The results from 10 datasets, recorded from five subjects, show that the hardware model is very accurate, with an average mean square error of 0.003. This suggests that the proposed design can mimic the behavior of clench force that a real limb does, and therefore this intelligent system could be a useful tool for any application related to prostheses.

A hybrid approach to plan itinerary for tourists

With the rapid development of information, communication and transportation systems, the number of tourists is increasing all over the world. Most of the tourists are facing problems when they visit unfamiliar places. Thats why it is often difficult for an individual tourist to make an optimal sightseeing itinerary plan during a tight schedule. To help tourists in this situation, a hybrid approach has been proposed in this paper to recommend an optimal itinerary plan. In this approach, K-means clustering algorithm has been used to cluster various tourist spots based on the geographic location and preferences of the tourists. Then the individual cluster is treated as a travelling salesman problem. To find the optimal route in each cluster greedy and 2-opt algorithms are applied. To switch from one cluster to the next cluster, a route is established from the last spot of one cluster to the centroid of the next cluster. Thus, all the tourist spots will be visited efficiently.

Using linear regression to forecast future trends in crime of Bangladesh

Crime is basically unpredictable and a social disturbance. With the increase in population of Bangladesh, the rate of crime is also increasing and affecting our society fatally in various ways. So it has become significant to analysis crime data for better understanding of future crime trends. In this case, machine learning and data mining techniques can play a significant role to discover future trends and patterns of crime. In this paper, linear regression model is used to forecast future crime trends of Bangladesh. The real dataset of crime is collected from the website of Bangladesh police. Then the linear regression model is trained on this dataset. After training the model, crime forecasting is done for dacoit, robbery, murder, women & child repression, kidnapping, burglary, theft and others for different region of Bangladesh. This work may be helpful for Bangladesh police and law enforcement agencies to forecast, prevent or solve future crime of Bangladesh.

Performance analysis of DS-CDMA under perfect and imperfect power control

CDMA refers to multiple access method in which the individual terminals uses spread spectrum techniques and occupy the entire spectrum whenever they transmit. This feature makes CDMA different from FDMA and TDMA. In the wireless communication, the Signal to Interference Ratio (SIR) and Bit Error Rate (BER) are the predominant parameter that characterizes the system performance. This paper presented here Standard Gaussian Approximation (SGA) methods presented in the international literature concerning the computation of the SIR and the BER in DS-CDMA systems under perfect and imperfect power control over fading and non-fading channel. The content and conclusions of this paper have driven to take many important decisions by varying different DS-CDMA communication parameters such as processing gain, number of interfering cells, multipath components etc. using SGA techniques. As SGA is analytically developed and is very computationally efficient solution for the system performance estimate in terms of SIR and BER, it helps us to avoid the tedious and cost-inefficient simulations.