Skip to main content
eScholarship
Open Access Publications from the University of California

Flexibility and Biases in Cognitive Control and Categorization

  • Author(s): Xu, Jing
  • Advisor(s): Ivry, Richard B
  • Griffiths, Thomas L
  • et al.
Abstract

How do we maintain balance in our daily life, yet at the same time, adapt to changes in the

environment? In the research presented here, I attempt to address this question in two ways. First,

I examine the flexibility and biases associated with processes of inhibitory control. Second, I

explore biases in categorization. Both lines of research are grounded in a probabilistic view of

the world.

In the first line of research, I examine two aspects of inhibitory control. In one study, I ask if

people can selectively inhibit their on-going actions. Using transcranial magnetic stimulation

(TMS) to probe corticomotor excitability, I examine if inhibition can be selectively directed to a

specific muscle, or if these inhibitory commands also modulate other muscles. In a second

experiment of this study, I trained participants to selectively stop. I develop a Sampling-Bias

model to analyze these data and find that costs associated with selective stopping are likely due

to a statistical sampling bias. In the second study on inhibitory control, I examine the effects of

intrinsic fluctuations in motor excitability on the dynamics of inhibitory control signals. Using

TMS applied prior to the start of a trial as a measure of excitability, I observed an interaction of

intrinsic and extrinsic effects, with the evidence suggesting that the dynamics of the control

signal outweigh the effects of intrinsic fluctuations.

In the second line of research, I use a novel experimental paradigm and computational model,

iterated-learning, to explore how biases influence categorization. In one study, I test the

prediction that iterated-learning converges to commonly shared biases, applying this to Bartlett's

classical serial-reproduction paradigm. In a second experiment, I use this paradigm to simulate

the cultural evolution of color categorization, seeking to understand if an iterated-learning

process, coupled with common perceptual and learning biases, will converge to the linguistic

universals found in the World Color Survey.

Both lines of research underscore a common theme: Understanding human cognition requires

consideration of our mental biases.

Main Content
Current View