This paper describes a method to reliably estimate latency of multifocal visual evoked potential (mfVEP) and a classifier to automatically separate reliable mfVEP traces from noisy traces. We also investigated which mfVEP peaks have reproducible latency across recording sessions. The proposed method performs cross-correlation between mfVEP traces and second order Gaussian wavelet kernels and measures the timing of the resulting peaks. These peak times offset by the wavelet kernel's peak time represents the mfVEP latency. The classifier algorithm performs an exhaustive series of leave-one-out classifications to find the champion mfVEP features which are most frequently selected to infer reliable traces from noisy traces. Monopolar mfVEP recording was performed on 10 subjects using the Accumap1™ system. Pattern-reversal protocol was used with 24 sectors and eccentricity upto 33°. A bipolar channel was recorded at midline with electrodes placed above and below the inion. The largest mfVEP peak and the immediate peak prior had the smallest latency variability across recording sessions, about ±2ms. The optimal classifier selected three champion features, namely, signal-to-noise ratio, the signal's peak magnitude response from 5 to 15Hz and the peak-to-peak amplitude of the trace between 70 and 250 ms. The classifier algorithm can separate reliable and noisy traces with a high success rate, typically 93%.
Download full-text PDF |
Source |
|---|---|
| http://dx.doi.org/10.1016/j.visres.2011.11.002 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
A lightweight model for echo trace detection in echograms based on improved YOLOv8.
Sci Rep
December 2024
College of Marine Living Resource Sciences and Management, Shanghai Ocean University, Shanghai, 201306, China.
Article Synopsis
- * Current detection models struggle with noise and complexity, making them hard to embed in echosounders effectively.
- * The paper presents YOLOv8-SBE, a streamlined fish detection model that improves efficiency, reduces parameters, and enhances small target recognition, achieving significant reductions in computational complexity and better accuracy metrics.
Panning for gold: Comparative analysis of cross-platform approaches for automated detection of political content in textual data.
PLoS One
November 2024
Institute for Communication Psychology and Media Education, Rheinland-Pfälzische Technische Universität Kaiserslautern-Landau, Landau, Germany.
To understand and measure political information consumption in the high-choice media environment, we need new methods to trace individual interactions with online content and novel techniques to analyse and detect politics-related information. In this paper, we report the results of a comparative analysis of the performance of automated content analysis techniques for detecting political content in the German language across different platforms. Using three validation datasets, we compare the performance of three groups of detection techniques relying on dictionaries, classic supervised machine learning, and deep learning.
View Article and Find Full Text PDFInvestigating Transfer Learning in Noisy Environments: A Study of Predecessor and Successor Features in Spatial Learning Using a T-Maze.
Sensors (Basel)
October 2024
Department of Physiology, Pusan National University School of Medicine, Yangsan 50612, Republic of Korea.
Article Synopsis
- This study examines how artificial agents perform in a noisy T-maze using Markov decision processes and specific learning algorithms (SF and PF).
- It assesses the impact of hyperparameters like reward learning rate (αr) and eligibility trace decay rate (λ) on the agents' adaptability, revealing that an αr of 0.9 consistently leads to better outcomes.
- The research highlights the importance of optimizing these hyperparameters for enhancing the effectiveness of learning algorithms, particularly in transfer learning contexts, and offers valuable insights for future work in AI and neuroscience.
Past rewinding of fluid dynamics from noisy observation via physics-informed neural computing.
Phys Rev E
August 2024
Department of Physics, Chung-Ang University, Seoul 06974, South Korea.
Reconstructing the past of observed fluids has been known as an ill-posed problem due to both numerical and physical challenges, especially when observations are distorted by inevitable noise, resolution limits, or unknown factors. When employing traditional differencing schemes to reconstruct the past, the computation often becomes highly unstable or diverges within a few backward time steps from the distorted and noisy observation. Although several techniques have been recently developed for inverse problems, such as adjoint solvers and supervised learning, they are also unrobust against errors in observation when there is time-reversed simulation.
View Article and Find Full Text PDFProtocol for neuron tracing and analysis of dendritic structures from noisy microscopy images using Neuronalyzer.
STAR Protoc
June 2024
Department of Biology, Technion - Israel Institute of Technology, Haifa 32000, Israel. Electronic address:
Article Synopsis
- The text introduces a protocol for using Neuronalyzer software to automatically reconstruct branched neuronal structures from noisy microscopy images.
- It outlines steps including loading images, denoising, segmenting, tracing, and extracting features like branch curvature and junction angles.
- The software is capable of batch processing and allows for statistical comparisons, making it versatile and effective in handling variations across different datasets.
Want AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!