Dynamic Neural Representations: An Inferential Challenge for fMRI

doi:10.1016/j.tics.2019.04.004

Trends in Cognitive Sciences

Volume 23, Issue 7, July 2019, Pages 534-536

https://doi.org/10.1016/j.tics.2019.04.004 Get rights and content

Measures of brain activity with high temporal resolution have shown that the information represented in a single brain region undergoes dynamic changes on the scale of milliseconds. This dynamic process presents a unique inferential challenge to low temporal resolution neural measures, such as BOLD fMRI. Potential solutions for fMRI requiring further investigation and development are discussed.

Section snippets

Acknowledgments

We thank Peter Bandettini for valuable discussion and comments. We gratefully acknowledge the support of the National Institute of Mental Health under R01 MH107797 (to A.G.) and ZIA MH002920–09 (to A.M.) and National Science Foundation under 1734907 (to A.G.).

References (12)

L.S. Glezer
Evidence for highly selective neuronal tuning to whole words in the “visual word form area”
Neuron
(2009)
A. Baeck
Influence of lexical status and orthographic similarity on the multi-voxel response of the visual word form area
Neuroimage
(2015)
F.H. Lin
Relative latency and temporal variability of hemodynamic responses at the human primary visual cortex
NeuroImage
(2018)
M. Misaki
Accurate decoding of sub-TR timing differences in stimulations of sub-voxel regions from multi-voxel response patterns
NeuroImage
(2013)
H. Liang
Interactions between feedback and lateral connections in the primary visual cortex
Proc. Natl. Acad. Sci. U. S. A.
(2017)
Y. Sugase
Global and fine information coded by single neurons in the temporal visual cortex
Nature
(1999)

There are more references available in the full text version of this article.

Cited by (19)

Enhancing neural encoding models for naturalistic perception with a multi-level integration of deep neural networks and cortical networks
2024, Science Bulletin
Cognitive neuroscience aims to develop computational models that can accurately predict and explain neural responses to sensory inputs in the cortex. Recent studies attempt to leverage the representation power of deep neural networks (DNNs) to predict the brain response and suggest a correspondence between artificial and biological neural networks in their feature representations. However, typical voxel-wise encoding models tend to rely on specific networks designed for computer vision tasks, leading to suboptimal brain-wide correspondence during cognitive tasks. To address this challenge, this work proposes a novel approach that upgrades voxel-wise encoding models through multi-level integration of features from DNNs and information from brain networks. Our approach combines DNN feature-level ensemble learning and brain atlas-level model integration, resulting in significant improvements in predicting whole-brain neural activity during naturalistic video perception. Furthermore, this multi-level integration framework enables a deeper understanding of the brain's neural representation mechanism, accurately predicting the neural response to complex visual concepts. We demonstrate that neural encoding models can be optimized by leveraging a framework that integrates both data-driven approaches and theoretical insights into the functional structure of the cortical networks.
From remembering to reconstruction: The transformative neural representation of episodic memory
2022, Progress in Neurobiology
Citation Excerpt :
As described above, perception includes a series of feedforward processing steps along a hierarchically organized neural pathway, during which different forms of neural representations are formed in corresponding regions. Through feedback processing (Kietzmann et al., 2019), the representations also change dynamically within single areas (Ghuman and Martin, 2019). For example, a nonhuman primate study showed that V1 neurons respond to visual orientation bars during early encoding but exhibit effects of contour integration 50–100 ms later that likely reflect feedback from V4 (Liang et al., 2017).
Although memory has long been recognized as a generative process, neural research of memory in recent decades has been predominantly influenced by Tulving’s “mental time traveling” perspective and focused on the reactivation and consolidation of encoded memory representations. With the development of multiple powerful analytical approaches to characterize the contents and formats of neural representations, recent studies are able to provide detailed examinations of the representations at various processing stages and have provided exciting new insights into the transformative nature of episodic memory. These studies have revealed the rapid, substantial, and continuous transformation of memory representation during the encoding, maintenance, consolidation, and retrieval of both single and multiple events, as well as event sequences. These transformations are characterized by the abstraction, integration, differentiation, and reorganization of memory representations, enabling the long-term retention and generalization of memory. These studies mark a significant shift in perspective from remembering to reconstruction, which might better reveal the nature of memory and its roles in supporting more effective learning, adaptive decision-making, and creative problem solving.
Latent brain state dynamics and cognitive flexibility in older adults
2022, Progress in Neurobiology
Citation Excerpt :
Human electrophysiological studies have revealed distinct activation and connectivity patterns associated with encoding, maintenance, and retrieval processes (Jensen and Lisman, 1998; Bashivan et al., 2014; Zakrzewska and Brzezicka, 2014), but the anatomical correlates have been hard to discern using scalp EEG recordings. In contrast, fMRI studies have typically not had the temporal resolution to disentangle brain responses associated with distinct phases (Ghuman and Martin, 2019; Glover, 2011), and, moreover, most previous studies have simply assumed that brain states are aligned with externally-driven task conditions and trial phases. We tested this assumption by examining how latent brain states emerge during working memory performance.
Cognitive impairment in older adults is a rapidly growing public health concern as the elderly population dramatically grows worldwide. While it is generally assumed that cognitive deficits in older adults are associated with reduced brain flexibility, quantitative evidence has been lacking. Here, we investigate brain flexibility in healthy older adults (ages 60–85) using a novel Bayesian switching dynamical system algorithm and ultrafast temporal resolution (TR = 490 ms) whole-brain fMRI data during performance of a Sternberg working memory task. We identify latent brain states and characterize their dynamic temporal properties, including state transitions, associated with encoding, maintenance, and retrieval. Crucially, we demonstrate that brain inflexibility is associated with slower and more fragmented transitions between latent brain states, and that brain inflexibility mediates the relation between age and cognitive inflexibility. Our study provides a novel neurocomputational framework for investigating latent dynamic circuit processes underlying brain flexibility and cognition in the context of aging.
Acute alcohol intoxication and expectations reshape the spatiotemporal functional architecture of executive control
2020, NeuroImage
Citation Excerpt :
However, the direction of the effect runs counter previous observations for decrease in ACC activity during Go/No-Go and Stroop tasks (Anderson et al., 2011; Marinkovic et al., 2012). Given the low temporal resolution of fMRI, the dynamic changes in single brain regions occurring within the millisecond range might have been mixed with other parallel activity (Ghuman and Martin, 2019). In addition, hemodynamic investigations of the effect of alcohol might be confounded by its vascular effects (Pike, 2012).
While the deleterious effects of acute ethyl alcohol intoxication on executive control are well-established, the underlying spatiotemporal brain mechanisms remain largely unresolved. In addition, since the effects of alcohol are noticeable to participants, isolating the effects of the substance from those related to expectations represents a major challenge. We addressed these issues using a double-blind, randomized, parallel, placebo-controlled experimental design comparing the behavioral and electrical neuroimaging acute effects of 0.6 vs 0.02 g/kg alcohol intake recorded in 65 healthy adults during an inhibitory control Go/NoGo task. Topographic ERP analyses of covariance with self-reported dose expectations allowed to dissociate their neurophysiological effects from those of the substance.
While alcohol intoxication increased response time variability and post-error slowing, bayesian analyses indicated that it did not modify commission error rates. Functionally, alcohol induced topographic ERP modulations over the periods of the stimulus-locked N2 and P3 components, arising from pre-supplementary motor and anterior cingulate areas. In contrast, alcohol decreased the strength of the response-locked anterior cingulate error-related component but not its topography. This pattern indicates that alcohol had a locally specific influence within the executive control network, but disrupted performance monitoring processes via global strength-based mechanisms.
We further revealed that alcohol-related expectations induced temporally specific functional modulations of the early N2 stimulus-locked medio-lateral prefrontal activity, a processing phase preceding those influenced by the actual alcohol intake.
Our collective findings thus not only reveal the mechanisms underlying alcohol-induced impairments in impulse control and error processing, but also dissociate substance- from expectations- related functional effects.
Phonemic segmentation of narrative speech in human cerebral cortex
2023, Nature Communications
THINGS-data, a multimodal collection of large-scale datasets for investigating object representations in human brain and behavior
2023, eLife

View all citing articles on Scopus

View full text

Trends in Cognitive Sciences

ForumDynamic Neural Representations: An Inferential Challenge for fMRI

Section snippets

Acknowledgments

Neuron

Neuroimage

NeuroImage

NeuroImage

Interactions between feedback and lateral connections in the primary visual cortex

Proc. Natl. Acad. Sci. U. S. A.

Global and fine information coded by single neurons in the temporal visual cortex

Nature

Forum
Dynamic Neural Representations: An Inferential Challenge for fMRI