The ECG classification is a critical task in the early and correct diagnosis of cardiovascular diseases. Although various models have been developed to tackle the heartbeat classification problem, their performance degrades on ECG signals recorded in varied testing conditions due to the distribution discrepancy among different sources of data. In this work, we have developed a multi-source domain generalization model to address the distribution discrepancy problem that occurred when the collection of the data is from multiple sources with various acquisition conditions. We have employed a combination of convolutional neural network (CNN) and long short term memory (LSTM) for feature extraction. Further, we exploit the adversarial domain generalization method to overcome probable heterogeneity between the train and test datasets. To increase generalization, we also utilized different augmentation techniques including random ECG pad and crop, adding low-frequency artifacts, and lead dropout. We evaluate our proposed model on cardiac abnormality classification based on 12-lead ECG signals associated with "Classification of 12-lead ECGs for the Phy-sioNet/Computing in Cardiology Challenge 2020". Our method, achieved a challenge validation score of 0.609, and full test score of 0.437 placing us (Sharif AI Team) 5th out of 41 teams in the final official ranking.

In this paper, application of Endurance Time (ET) method in nonlinear seismic analysis of o shore pile supported systems has been studied. The ET method is a time-history analysis in which structures are subjected to intensifying arti cial acceleration functions. The ET method reduces complexity and computational demand of conventional nonlinear seismic analysis, and it provides response at di erent seismic levels in a single ET analysis. The aforementioned methodology has been applied to a typical model of single pile and then to a functional jacket o shore platform in Persian Gulf region. Seismic response of aforesaid models by ET method has been compared with conventional time-history method. The results indicate that ET method is reliable in capturing seismic response of o shore platforms supported on piles with an acceptable accuracy.

Background High‐level cognitive processes such as binding the processed data from sensory modules with elements stored in the memory involve the activity of long‐range pyramidal cells. These excitatory neuronal populations also provide input to a population of GABAAergic inhibitory interneurons, which in turn recruit feedback links to suppress the activity of the pyramidal cells. Interneuron networks generate rhythmic synchronization in the Gamma band driven by the time constant of the GABA receptors. Disrupted or desynchronized Gamma oscillations have been observed in patients of Alzheimer’s disease (AD). Earlier works have proposed the deficit in coherence between oscillations measured by EEG electrodes across the frontal lobe in the Gamma band in response to olfactory stimulation as a diagnostic marker of AD. This study examines the strength and spatial spread of Gamma band activity induced by auditory chirp stimulation as a marker for AD. The chirp signal is designed to entrain a target frequency of 40Hz at which the populations of inhibitory interneurons are known to operate. Method A session comprising 11 interleaved periods of 40sec ON and 20sec OFF auditory stimuli of 5kHz tone modulated by a 40Hz chirp at 0.1 duty cycle was administered to mild AD patients and non‐AD elderly participants with memory complaints, and EEG data were collected by a 10/20 system. Magnitude of 40Hz oscillations at different scalp positions during the ON cycles was measured as an indicator of the entrained Gamma oscillations. Result While 40Hz oscillations were recorded across a majority of electrodes in non‐AD demented participants with particular strengths in the temporal and frontal areas, the 40Hz entrainment occurred for a limited number of electrodes in AD patients. Conclusion Auditory chirp stimulation at 40Hz results in spatially distinguishable patterns of entrained Gamma oscillations in AD patients and non‐AD demented participants, and hence suggests a marker for AD. Despite this difference, the fact that 40Hz entrainment still occurs in regions of the brain in AD patients offers a positive indication for the possibility to employ such stimulation to reinvigorate the operation of the involved neural circuitry in therapy campaigns. Further studies are needed to assess such possibility.

No abstract