Biology subjects

Cai, M. B.

Publications and source records attributed to Cai, M. B..

3 recordsLinked to original sources

bioRxiv2024

Geometry representations along visual pathways in human spatial navigation

The representation of geometric structures of ones surroundings is key to self-localization during human spatial navigation. However, the spatial organization of geometry representation in the visual system has not been fully characterized. By modeling the synchronized brain activity from participants watching navigational videos from identical realistic virtual environments under different weather and lighting conditions, we found a compact representation of egocentric 3D scene geometric structures is present in a widespread network of brain regions. These regions carrying geometry-related signals form three parallel pathways, which we collectively refer to as "geometry visual pathways", starting from the primary visual cortex: the dorsal and medial pathways end in the intraparietal areas, while the ventral pathway arrives at the hippocampus via the parahippocampal gyrus. The synchronized neural activity in the identified geometry visual pathways allows for reliable decoding of the 3D scene geometric structures. Furthermore, road types, a more abstract representation of the route geometry, are encoded in overlapping pathways, distinctly absent from early visual cortex (V1, V2, V3). In addition to complementing the classical "what" and "where" dichotomy, the identified brain-wide geometry visual pathways narrow a critical gap in understanding how the brain constructs cognitive maps from visual inputs.

neuroscience↗

bioRxiv2019

Facilitating open-science with realistic fMRI simulation: validation and application

BackgroundWith advances in methods for collecting and analyzing fMRI data, there is a concurrent need to understand how to reliably evaluate and optimally use these methods. Simulations of fMRI data can aid in both the evaluation of complex designs and the analysis of data. New MethodWe present fmrisim, a new Python package for standardized, realistic simulation of fMRI data. This package is part of BrainIAK: a recently released open-source Python toolbox for advanced neuroimaging analyses. We describe how to use fmrisim to extract noise properties from real fMRI data and then create a synthetic dataset with matched noise properties and a user-specified signal. ResultsWe validate the noise generated by fmrisim to show that it can approximate the noise properties of real data. We further show how fmrisim can help researchers find the optimal design in terms of power. Comparison with other methodsfmrisim ports the functionality of other packages to the Python platform while extending what is available in order to make it seamless to simulate realistic fMRI data. ConclusionsThe fmrisim package holds promise for improving the design of fMRI experiments, which may facilitate both the pre-registration of such experiments as well as the analysis of fMRI data. HighlightsO_LIfmrisim can simulate fMRI data matched to the noise properties of real fMRI. C_LIO_LIThis can help researchers investigate the power of their fMRI designs. C_LIO_LIThis also facilitates open science by making it easy to pre-register analysis pipelines. C_LI

neuroscience↗

bioRxiv2018

Representational structure or task structure? Bias in neural representational similarity analysis and a Bayesian method for reducing bias

The activity of neural populations in the brains of humans and animals can exhibit vastly different spatial patterns when faced with different tasks or environmental stimuli. The degree of similarity between these neural activity patterns in response to different events is used to characterize the representational structure of cognitive states in a neural population. The dominant methods of investigating this similarity structure first estimate neural activity patterns from noisy neural imaging data using linear regression, and then examine the similarity between the estimated patterns. Here, we show that this approach introduces spurious bias structure in the resulting similarity matrix, in particular when applied to fMRI data. This problem is especially severe when the signal-to-noise ratio is low and in cases where experimental conditions cannot be fully randomized in a task. We propose Bayesian Representational Similarity Analysis (BRSA), an alternative method for computing representational similarity, in which we treat the covariance structure of neural activity patterns as a hyper-parameter in a generative model of the neural data. By marginalizing over the unknown activity patterns, we can directly estimate this covariance structure from imaging data. This method offers significant reductions in bias and allows estimation of neural representational similarity with previously unattained levels of precision at low signal-to-noise ratio. The probabilistic framework allows for jointly analyzing data from a group of participants. The method can also simultaneously estimate a signal-to-noise ratio map that shows where the learned representational structure is supported more strongly. Both this map and the learned covariance matrix can be used as a structured prior for maximum a posteriori estimation of neural activity patterns, which can be further used for fMRI decoding. We make our tool freely available in Brain Imaging Analysis Kit (BrainIAK).\n\nAuthor summaryWe show the severity of the bias introduced when performing representational similarity analysis (RSA) based on neural activity pattern estimated within imaging runs. Our Bayesian RSA method significantly reduces the bias and can learn a shared representational structure across multiple participants. We also demonstrate its extension as a new multi-class decoding tool.

neuroscience↗