A5046525447- Image and Signal Denoising Methods
- Neural dynamics and brain function
- Visual perception and processing mechanisms
- EEG and Brain-Computer Interfaces
- Species Distribution and Climate Change
- Bioinformatics and Genomic Networks
- Neuroscience and Neural Engineering
- Image Processing and 3D Reconstruction
- Gaze Tracking and Assistive Technology
- Image and Object Detection Techniques
- Medical Image Segmentation Techniques
- Social Media in Health Education
- Mental Health Research Topics
- Face Recognition and Perception
- Advanced Image Fusion Techniques
- Functional Brain Connectivity Studies
- Anomaly Detection Techniques and Applications
- Parkinson's Disease Mechanisms and Treatments
- Smart Agriculture and AI
Yale University
2021-2022
Witten/Herdecke University
2022
Columbia University
2017-2019
University of Colorado Boulder
1999
This paper describes a novel hierarchical system for shared control of humanoid robot. Our framework uses low-bandwidth Brain Computer Interface (BCI) to interpret electroencephalography (EEG) signals via Steady-State Visual Evoked Potentials (SSVEP). BCI allows user reliably interact with the humanoid. clearly delineates between autonomous robot operation and human-guided intervention control. shared-control leverages ability accomplish low level tasks on its own, while assists high...
Clinical scores (disease rating scales) are ordinal in nature. Longitudinal studies which use clinical produce time series. These series tend to be noisy and often have a short-duration. This paper proposes denoising method for such The uses hierarchical approach draw statistical power from the entire population of study's patients give reliable, subject-specific results. is applied MDS-UPDRS motor Parkinson's disease.
The hierarchical architecture of deep convolutional neural networks (CNN) resembles the multi-level processing stages human visual system during object recognition. Converging evidence suggests that this organization is key to CNN achieving human-level performance in categorization [22]. In paper, we leverage investigate spatiotemporal dynamics rapid brain. Specifically focus on perceptual decisions associated with different levels ambiguity. Using simultaneous EEG-fMRI, demonstrate temporal...
Pharmaceutical companies increasingly must consider patients' needs in drug development.Since are often difficult to measure, especially rare diseases, information development decision-making is limited.In the proposed study, we employ opportunity algorithm identify and prioritize unmet medical of multiple sclerosis patients shared social media posts.Using topic modeling sentiment analysis features generated.The result implies that sensory problems, pain, mental health fatigue sleep...
Virtual reality (VR) offers the potential to study brain function in complex, ecologically realistic environments. However, additional degrees of freedom make analysis more challenging, particularly with respect evoked neural responses. In this paper we designed a target detection task VR where varied visual angle targets as subjects moved through three dimensional maze. We investigated how latency and shape classic P300 response locking electroencephalogram data image onset, target-saccade...
The digitization of natural history collections over the past three decades has unlocked a treasure trove specimen imagery and metadata. There is great interest in making this data more useful by further labeling it with additional trait data, modern deep learning machine techniques utilizing convolutional neural nets (CNNs) similar networks show particular promise to reduce amount required manual human experts, process much faster less expensive. However, most cases, accuracy these...
There is much excitement across a broad range of biological disciplines over the prospect using deep learning and similar modern statistical methods to label research data. The extensive time, effort, cost required for humans dataset drastically limits type amount data that can be reasonably utilized, currently major bottleneck application datasets such as specimen imagery, video audio recordings. While number researchers have shown how convolutional neural networks (CNN) trained classify...
The Hamilton Depression Rating Scale provides ordinal ratings for evaluating different aspects of depression. These are usually quite noisy, and longitudinal patterns in the can be difficult to discern. This paper proposes a hierarchical maximum-a-posteriori (MAP) method denoising time series such ratings. Real-world data from clinical trial analyzed using model. Denoising reveals subject-specific patterns, predicts future ratings, progression via principal component analysis.