Biology subjects

Zhu, Z.-C.

Publications and source records attributed to Zhu, Z.-C..

3 recordsLinked to original sources

bioRxiv2022

Neural Representations of Self-Generated Thought during Think-aloud fMRI

Is the brain at rest during the so-called resting-state? Ongoing experiences in the resting-state vary in unobserved and uncontrolled ways across time, individuals, and populations. However, the role of self-generated thoughts in resting-state fMRI remains largely unexplored. In this study, we collected real-time self-generated thoughts during "resting-state" fMRI scans via the think-aloud method (i.e., think-aloud fMRI), which required participants to report whatever they were currently thinking. We first investigated brain activation patterns during a think-aloud condition and found that significantly activated brain areas included all brain regions required for speech. We then calculated the relationship between divergence in thought content and brain activation during think-aloud and found that divergence in thought content was associated with many brain regions. Finally, we explored the neural representation of self-generated thoughts by performing representational similarity analysis (RSA) at three neural scales: a voxel-wise whole-brain searchlight level; a region-level whole-brain analysis using the Schaefer 400-parcels; at the systems level using the Yeo seven-networks. We found that "resting-state" self-generated thoughts were distributed across a wide range of brain regions, involving all seven Yeo networks. This study highlights the value of considering ongoing experiences during resting-state fMRI, as well as providing preliminary methodological support for think-aloud fMRI.

neuroscience↗

bioRxiv2020

An Industrial-Grade Brain Imaging-Based Deep Learning Classifier

Beyond detecting brain lesions or tumors, comparatively little success has been attained in identifying brain disorders such as Alzheimers disease (AD), based on magnetic resonance imaging (MRI). Many machine learning algorithms to detect AD have been trained using limited training data, meaning they often generalize poorly when applied to scans from previously unseen populations. Therefore, we built a practical brain MRI-based AD diagnostic classifier using deep learning/transfer learning on a dataset of unprecedented size and diversity. A retrospective MRI dataset pooled from more than 217 sites/scanners constituted one of the largest brain MRI samples to date (85,721 scans from 50,876 participants) between January 2017 and August 2021. Next, a state-of-the-art deep convolutional neural network, Inception-ResNet-V2, was built as a sex classifier with high generalization capability. The sex classifier achieved 94.9% accuracy and served as a base model in transfer learning for the objective diagnosis of AD. After transfer learning, the model fine-tuned for AD classification achieved 90.9% accuracy in leave-sites-out cross-validation on the Alzheimers Disease Neuroimaging Initiative (ADNI, 6,857 samples) dataset and 94.5%/93.6%/91.1% accuracy for direct tests on three unseen independent datasets (AIBL, 669 samples / MIRIAD, 644 samples / OASIS, 1,123 samples). When this AD classifier was tested on brain images from unseen mild cognitive impairment (MCI) patients, MCI patients who converted to AD were 3 times more likely to be predicted as AD than MCI patients who did not convert (65.2% vs 20.6%). Predicted scores from the AD classifier showed significant correlations with illness severity. In sum, the proposed AD classifier offers a medical-grade marker that has potential to be integrated into AD diagnostic practice.

neuroscience↗

bioRxiv2019

Reproducible observations revealing patterns of the default mode network underlie rumination

Rumination is a repetitive self-referential thinking style and posited to be an expression of abnormalities in the default mode network (DMN) in major depressive disorder (MDD). Recent evidences indicate DMN is not a unitary network but can be further divided into 3 functionally heterogenous subsystems. However, the subsystem mechanism through which DMN underlie rumination remain unclear. Here, with a modified continuous state-like paradigm, we induced healthy participants to ruminate or imagine objective scenarios (as a distraction control condition) on 3 different MRI scanners. We compared functional connectivities (FC) and inter-subject correlations (ISC) of the DMN and its 3 subsystems between rumination and distraction states. Results yielded a highly reproducible and dissociated pattern. During rumination, within-DMN FC was generally decreased compared to the distraction state. At the subsystem level, we found increased FC between the core and medial temporal lobe (MTL) subsystem and decreased FC between the core and dorsal medial prefrontal cortex (DMPFC) subsystem and within the MTL subsystem. Furthermore, we found decreased ISC within the MTL subsystem. These results suggest a specific and less synchronous activity pattern of DMN during rumination and shed new light on the association between rumination and DMN substrates regarding MDD.

neuroscience↗