Skip to search formSkip to main contentSkip to account menu

Young Children Bet on Their Numerical Skills

It was shown that children as young as 5 years made metacognitive “bets” on their numerical discriminations in a wagering task, and metacognition ability in only the numerical domain predicted their school-based mathematics knowledge.
View on SAGE (opens in a new tab)

Dissociable signatures of visual salience and behavioral relevance across attentional priority maps in human cortex

This work tested the hypothesis that visual salience and behavioral relevance independently impact the activation profile across retinotopically-organized cortical regions by quantifying attentional priority maps measured in human brains using functional MRI while participants attended one of two differentially-salient stimuli.
View on PubMed (opens in a new tab)

Spatial Tuning Shifts Increase the Discriminability and Fidelity of Population Codes in Visual Cortex

It is found that vRF position shifts contribute more to population-level enhancements of visual information than changes in vRF size or gain, which suggests that position shifts are a principal mechanism by which spatial attention enhances population codes for relevant visual information.
View on Publisher (opens in a new tab)

Interpretable multi-timescale models for predicting fMRI responses to continuous natural speech

This work constructs interpretable multi-timescale representations by forcing individual units in an LSTM LM to integrate information over specific temporal scales, which allows us to explicitly and directly map the timescale of information encoded by each individual fMRI voxel.
View via Publisher (opens in a new tab)

Inverted Encoding Models Assay Population-Level Stimulus Representations, Not Single-Unit Neural Tuning

It is argued that using stimulus reconstructions to infer properties of single neurons, such as neural tuning bandwidth, is an ill-posed problem with no unambiguous solution.
View on Publisher (opens in a new tab)

Multi-timescale representation learning in LSTM Language Models

This work constructs explicitly multi-timescale language models by manipulating the input and forget gate biases in a long short-term memory (LSTM) network and empirically analyze the timescale of information routed through each part of the model using word ablation experiments and forgetting gate visualizations.
View PDF on arXiv (opens in a new tab)

Low-Dimensional Structure in the Space of Language Representations is Reflected in Brain Responses

An encoder-decoder transfer learning method from computer vision is adapted to investigate the structure among 100 different feature spaces extracted from hidden representations of various networks trained on language tasks to reveal a low-dimensional structure where language models and translation models smoothly interpolate between word embeddings, syntactic and semantic tasks, and future word embedDings.
View PDF on arXiv (opens in a new tab)

Multivariate Analysis of BOLD Activation Patterns Recovers Graded Depth Representations in Human Visual and Parietal Cortex

This work recorded BOLD fMRI while human subjects viewed a stereoscopic random-dot sphere at various positions along the depth and horizontal axes, and the stimuli were presented across a wider range of disparities compared to previous neuroimaging studies, allowing for validated method of recovering depth representations from retinotopic cortex.
View on Publisher (opens in a new tab)

Value-driven attentional capture enhances distractor representations in early visual cortex

It is suggested that value-driven attentional capture begins with sensory modulations of distractor representations in early areas of visual cortex, and the fidelity of neural representations related to task-irrelevant distractors increased when the distractors were previously associated with a high reward.
View on PubMed (opens in a new tab)

Approximating Stacked and Bidirectional Recurrent Architectures with the Delayed Recurrent Neural Network

The delayed-RNN is explored, which is a single-layer RNN that has a delay between the input and output, and it is proved that a weight-constrained version of the delayed- RNN is equivalent to a stacked RNN, and that the delay gives rise to partial acausality, much like bidirectional networks.
View Paper (opens in a new tab)
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy (opens in a new tab), Terms of Service (opens in a new tab), and Dataset License (opens in a new tab)