About the lab

http://www.ulbertlab.com

- Development and testing of MEMS-based high-denstity neural sensors
- Investigation of neural oscillations in animal and human brain tissue
- EEG-based (non-invasive) brain-computer interface devices for human use
- Integration of opto-electrical recording techniques both in vitro and in vivo

Featured projects (1)

Featured research (6)

Objective: The growing number of recording sites of silicon-based probes means that an increasing amount of neural cell activities can be recorded simultaneously, facilitating the investigation of underlying complex neural dynamics. In order to overcome the challenges generated by the increasing number of channels, highly automated signal processing tools are needed. Our goal was to build a spike sorting model that can perform as well as offline solutions while maintaining high efficiency, enabling high-performance online sorting. Approach: In this paper we present ELVISort, a deep learning method that combines the detection and clustering of different action potentials in an end-to-end fashion. Main results: The performance of ELVISort is comparable with other spike sorting methods that use manual or semi-manual techniques, while exceeding the methods which use an automatic approach: ELVISort has been tested on three independent datasets and yielded average F1 scores of 0.96, 0.82 and 0.81, which comparable with the results of state-of-the-art algorithms on the same data. We show that despite the good performance, ELVISort is capable to process data in real-time: the time it needs to execute the necessary computations for a sample of given length is only 1/15.71 of its actual duration (i.e. the sampling time multipled by the number of the sampling points). Significance: ELVISort, because of its end-to-end nature, can exploit the massively parallel processing capabilities of GPUs via deep learning frameworks by processing multiple batches in parallel, with the potential to be used on other cutting-edge AI-specific hardware such as TPUs, enabling the development of integrated, portable and real-time spike sorting systems with similar performance to offline sorters.
Eyetracking opens new possibilities for people with physical disabilities, enabling them to explore new ways of communication and interaction with their environment. However, this field of assistive technology is not available to everyone in need and still faces substantial obstacles: high-cost devices for use in real-world environments, difficulty in enhancing the performance of a calibrated system, difficulty in providing a personalized, adaptive setup. The aim of this work is to provide a lightweight, low cost, wireless device with a software performing real-time 2-D gaze prediction in a fixed coordinate system, and the possibility of enhancing performance by recalibration.
Classification of electroencephalography (EEG) signals is a fundamental issue of Brain Computer Interface (BCI) systems, and deep learning techniques are still under investigation although they are dominant in other fields like computer vision and natural language processing. In this paper, we introduce the chessboard image transformation method in which the motor imagery EEG signals were transformed into images in order to be classified using a hybrid deep learning model. The EEG motor movement/imagery Physionet dataset was used and the Motor Imagery (MI) signals for two frequency bands (Mu [8–13 Hz] and Beta [13–30 Hz]) were transformed into 2-channel images (one channel for each band). The network model consists of Deep Convolutional Neural Network (DCNN) to extract the spatial and frequency features followed by Long Short Term Memory (LSTM) to extract temporal features and then finally to be classified into 5 different classes (4 motor imagery tasks and one rest). The results were promising with 68.72% classification accuracy for the chessboard approach compared to 68.13% for the azimuthal projection with Clough-Tocher interpolation (2-bands scenario) and to 64.64% average accuracy for a baseline method, i.e., Support Vector Machine (SVM).

Lab head

István Ulbert
Department
  • Institute of Cognitive Neuroscience and Psychology

Members (14)

Lucia Wittner
  • Research Center for Natural Sciences, Budapest, Hungary
Richárd Fiáth
  • Research Centre for Natural Sciences, Hungarian Academy of Sciences, Budapest
Bálint File
  • Pázmány Péter Catholic University
Tibor Nánási
  • Alkahest, Inc.
Gergely Márton
  • Research Centre for Natural Sciences
Domokos Meszéna
  • Institute of Cognitive Neuroscience
Kinga Toth
  • Hungarian Academy of Sciences
Domonkos Horváth
  • Hungarian Academy of Sciences
Estilla Zsófia Tóth
Estilla Zsófia Tóth
  • Not confirmed yet
Erick Noboa
Erick Noboa
  • Not confirmed yet

Alumni (4)

Katharina T Hofer
  • Hebrew University of Jerusalem
Ildikó Pál
  • Gedeon Richter Plc
Zoltan Karasz
  • Pázmány Péter Catholic University
Bálint Péter Kerekes
  • Hungarian Academy of Sciences