Beneficial Perturbation Network for Lifelong Learning

Communicated by Dr. Yuxiao Yang

RESEARCH

December 2021

Shixian Wen, Laurent Itti

Lifelong learning challenges

The human brain can quickly learn and adapt its behavior in a wide range of environments throughout its lifetime. In contrast, deep neural networks only learn one sophisticated but fixed mapping from inputs to outputs.  In more complex and dynamic scenarios where the inputs to outputs mapping may change with different contexts, the deployment of these deep neural network systems would be constrained. One of the failed salient scenarios is lifelong learning—learning new independent tasks sequentially without forgetting previous tasks. More specifically, agents should incrementally learn and evolve based on multiple tasks from various data distributions across time while remembering previously learned knowledge. In general, current neural networks are not capable of lifelong learning and usually suffer from “catastrophic forgetting”—learning the knowledge of the new task would overwrite the fixed learned mapping of an old task. This effect typically leads to a significant decrease of the network performances on previous tasks or, in the worst case, leads to the network completely forgetting all previous tasks.

Brain-inspired cognitive model with attention for self-driving cars

Communicated by Dr. Sung-Phil Kim

RESEARCH

December 2021

Shitao Chen, Songyi Zhang, Badong Chen and Nanning Zheng

As a typical artificial intelligence system, self-driving cars, unlike normal artificial intelligence systems, usually concern the safety of people’s lives and property, and have little tolerance of mistakes. With the furthering of research on self-driving technology, the existing computing framework based on the “perception-planning-decision-control” information processing method has increasingly manifested the problems of low computing efficiency, poor environmental adaptability, and insufficient self-learning ability. Our research work mainly refers to the psychological level of human cognition to construct a new type of self-driving method.

Cross-Cultural Exploration of Neuroethics in Engineering

EVENT

December 2021

Cynthia Weber, PhD, on behalf of IEEE Brain

Guidelines that consider societal and cultural impacts of neurotechnology are crucial for ensuring responsible innovation in the field.

Ethical considerations have not always been of primary concern in the development of technology. However, the need for ethical standards and guidelines for neurotechnology has received significant support with multiple efforts underway that aim to sidestep past mistakes by preparing for future development and use cases. The challenge lies in identifying the complex social, legal, and cultural issues tied to how neurotechnologies will be accessed and implemented once released into the world, and the associated safety, privacy, and long-term consequences of its use. For many people, the brain is intimately connected to one’s sense of self and personal identity—our thoughts and emotions, for example. Consequently, neurotechnology devices that intervene with the brain, whether for medical treatment, wellness applications, or entertainment, may pose unique perceived risks for the user. This is also the case when neurotechnology has the potential to be implemented in employment, legal, or educational contexts. In all these scenarios, ethical considerations are interwoven within layers of consent, data access and control, and possible manipulation.

Message from the Editor

OPINION

September 2020

Ricardo Chavarriaga

The COVID19 pandemic has brought upon us an unprecedented situation. The current events have had a profound impact at a global scale and have once more shown the important role that science and technology play in providing solutions to societal challenges.

Temporary Tattoo Electrodes for brain recordings in clinical settings

RESEARCH

September 2020

Laura M. Ferrari1*, E. Ismailova3*, F.Greco2*

Temporary Tattoo Electrodes (TTEs) are dry and conformable electrodes that are able to capture weak surface electrophysiological signals while being imperceptible for the user. We demonstrated the use and characterised TTEs in a clinical electroencephalography (EEG) monitoring set-up, proving for the first time the compatibility of a dry electrode with magnetoencephalography (MEG) sensors (Figure 1) (1).

Bayesian optimization for automated neurostimulation: future directions and challenges

RESEARCH

September 2020

Samuel Laferrière1,2, Marco Bonizzato3, Numa Dancause3, & Guillaume Lajoie1,4

The stimulation optimization problem & the rapid evolution of electrode technology:
The development of neurostimulation techniques for targeted biomarker control is an active area of research. New implantable devices are microfabricated with hundreds or thousands of electrodes, holding great potential for precise spatiotemporal stimulation. These interfaces not only serve as a crucial experimental tool to probe computation in neural circuits [7,8,9], but also have applications in neuroprostheses used to aid recovery of motor, sensory and cognitive modalities affected by injury or disease [14-19]. Yet, existing electrical neuromodulation interventions do not fully take advantage of the rich stimulation repertoire advanced electrode technologies offer, instead relying mostly on incomplete and manual input-output mapping, and often on single electrode stimulation [1,6].

What large-scale analysis tells us about EEG pre-processing

RESEARCH

September 2020

Kay Robbins1 Senior Member, IEEE and Tim Mullen2, Member IEEE

Although electroencephalography (EEG) is an important high time-resolution brain imaging technology used in laboratory, clinical, and even consumer applications, consistent handling of signal artifacts continues to be an important challenge. In a recent series of papers [1] [2] [3], we and collaborators compared EEG analysis results across multiple studies, EEG headset types, and preprocessing methods. We considered channel and source signal characteristics and explored time-locked event analysis. The work produced several insights of general interest to EEG researchers, as outlined below.

Book Review – Neural Engineering, 3rd Edition (Bin He, Editor)

BOOK REVIEW

May 2020

Bruce Wheeler, PhD

Dr. Bin He is to be congratulated on pulling together an even stronger set of contributors and topics to make the third edition of Neural Engineering (Bin He, editor; Springer) a significant enhancement over the second edition. Easiest to note are the inclusion of 22 chapters (an increase of 3), with nine new topics, and three previous topics presented by new authors with fresh perspectives. Perhaps over half the material is new. A quick additional look shows that the new topics are quite timely.

Transfer Learning for Brain-Computer Interfaces: Euclidean Alignment and Label Alignment

RESEARCH

May 2020

Dongrui Wu and He He

Ministry of Education Key Laboratory of Image Processing and Intelligent Control, School of Artificial Intelligence and Automation, Huazhong University of Science and Technology, Wuhan, China.
Email: drwu@hust.edu.cn, hehe91@hust.edu.cn.

A brain-computer interface (BCI) system [1], [2] acquires the brain signal, decodes it, and then translates it into control commands for external devices, so that a user can interact with his/her surroundings using thoughts directly.

Re-designing the Wheel: The High Relevance of EEG in Studying Brain Networks

RESEARCH

May 2020

Abbas Sohrabpour and Bin He

There seem to be two major principles that govern brain function; functional segregation and functional integration [1]. The brain is a highly specialized, and at the same time, a highly integrated organ. Spatially segregated regions are tuned to perform special functions optimally (functional segregation), and at a higher level, multiple regions need to pull resources together, and integrate functions, to perform complex tasks (functional integration).