Kazuyuki Samejima

Ph.D
Professor

Tamagawa University
Brain Science Institute
5-1-1 Tamagawa gakuen, Machida, Tokyo
194-8610

Tel: 042-739-8668
Fax: 042-739-8663
samejima[at]lab.tamagawa.ac.jp

Bio

Ph. D. in Engineering from Tokyo University of Agriculture and Technology in 1999
1999 Researcher, Japan Science and Technology Agency (JST)
2003 Researcher, ATR Brain Information Institute
2005 Lecturer, Tamagawa University Research Institute
2007 Associate Professor, Tamagawa University
2016 to now, Professor, Brain Science Institute , Tamagawa University

Research area
machine learning with a focus on reinforcement learning, computational neuroscience, and neurophysiology of the basal ganglia.

A board member of the Japanese Neural Network Society. A member of the Japanese Society for Animal Psychology, the Japanese Neuroscience Society, the Society for Neuroscience, the Institute of Electronics, Information and Communication Engineers, and the Japanese Society for Human Behavior and Evolution.

Research Interests

Neural mechanisms of decision making and learning

We are studying the neural mechanisms of how our decision-making is processed in the brain. How do we decide what to do? The goodness or badness of the outcome after a choice adapts our behavior better. But where and how does the brain do that? How do we accumulate experience? How are situational judgments made? What do we pursue as rewards? What conflicts do we face in our interactions with others?
To understand the mechanisms of decision-making, we are challenging these questions not only through experimental methods such as neurophysiology, non-invasive brain activity measurement techniques, and cognitive psychology/behavioral experiments, but also by combining them with computational theories that provide mathematical models of decision-making, such as artificial intelligence, economics, psychology, and behavioral ecology.

Research Topics

1.Value functions and representation learning in the basal ganglia

Electrophysiological and brain imaging studies have shown that the basal ganglia play an important role in reward-based decision making. We are studying how neural activities are represented during decision making at the level of individual neurons. Previous studies have shown that information about the value of predicting future rewards is represented in each choice, and that information about the value of chosen target before choosing a movement is represented in neuronal activity in the striatum of the basal ganglia.
We use electrophysiological methods to measure neural activity in monkeys during behavior and the theory of reinforcement learning to study neural representations and algorithms. How are neural circuits in the cerebral cortex and basal ganglia involved in "choice" to make use of past experiences in the future? My goal is to elucidate the mechanism of decision making, such as how information processing is abnormal in neurological diseases such as addiction.

2.Mechanisms for Generating Communication Symbols

Our actions are not only to move around in our environment or manipulate objects, but also to communicate with others. Using symbols to communicate our intentions to others is one of our essential abilities to cooperate with others. But how do we learn, agree with others, and use it as communicative language?
To understand the neural mechanisms that allow us to communicate with others in written and spoken language, we are using an experimental paradigm in which subjects form agreements with the meaning of symbols and cooperate with others. These learning behaviors in the paradigm could be modeled by using Bayesian prediction framework and reinforcement learning. This will help to reveal the evolutionary nature of language learning and the innate biases that humans have. Our goal is to clarify the basic principles of language communication and the nature of learning.

3.Psychological and neural basis of human-animal communication

People also communicate using non-verbal information. Through gestures, nods, gazes, facial expressions, etc., we can infer the cognitive state (what we understand or not) and emotional state (pleasant, unpleasant, happy, sad, etc.) of others, and act accordingly. To understand this mechanism, we can use communication between animals and humans. It is convenient to study about non-verbal communication because animals are not capable of verbal understanding . Domesticated animals, such as dogs and horses, have been shown to be able to use human emotional states and gaze information. Nonverbal communication becomes more effective as the bond is formed. Our goal is to clarify the learning process psychologically and behaviorally, and to elucidate the neural mechanisms by capturing the corresponding changes in the brain.

People

Samejima Kazuyuki

Professor

Mari Kumashiro

Researcher

Taiji Yamada

PhD Candidate

Takeru Kamioka

Master Course Student

Ayami Suzuki

Lab Manager

鮫島研究室メンバー集合写真
In front of the HBSH building, taken 04 Apr 2023.

Get Involved

​​Thank you for your interest in participating with the Samejima Lab. We are currently recruiting Master and PhD candidates and also Post-doc researchers of JSPS.
If you are interested in learning more about participating in our research, please get in touch with us via contact form below with your contact information.

Contact Us

Please contact us if you have any questions about our work. We look forward to hearing from you.






Email Address

samejima[at]lab.tamagawa.ac.jp

Telephone Number

+81-42-739-8668

Address

Tamagawa University
Brain Science Institute
6-1-1 Tamagawa gakuen, Machida Tokyo
194-8610

Directions

The Samejima Lab is located in the Brain Science Institute at Tamagawa University (Three-minute walk from Tamagawagakuen-Mae station on the Odakyu Odawara Line).