Neuroscience preprint round-up
Our AI writes about the latest Neuroscience preprints published on bioRxiv
The individual reports below - including each headline - were generated automatically by our machine-reading software from an RSS feed of bioRxiv Neuroscience preprints on 10 March 2021.
How does it work? Briefly, the technology comprises an ensemble of language models fine-tuned for extractive and abstractive summarisation built with spaCy, huggingface transformers and Scholarcy technology.
The headlines in italic are generated by our abstractive engine. The longer synopses use our extractive engine, and consist of automatically extracted highlights from the original texts, with some subclause truncation, transitional phrase removal, and rewriting in the third person.
Young adults have limited access to their own facial expressions.
A research group led by A. Ciston of the Department of Psychology (2021) studied limited metacognitive access to one's own facial expressions. The analyses suggest that healthy young volunteers were only able to estimate their performance in producing non-stereotypical facial expressions based on partial information. This indicates that they do not have metacognitive access to the low-level details of the facial expressions. They speculate that this distinction might be related to the lack of concurrent visual information during social interactions. This is surprising, they argue, because it sets facial movements apart from other body movements.
40 healthy participants were included in the study. The authors’ findings appear to reinforce prior research in this field: “To examine this possibility, we built linear regression ML models on confidence ratings that included differences for each landmark as individual features. This analysis revealed that the models built for all participants could predict confidence from the combined features,” Ciston suggested. However, “In our exploratory analyses, we found a clear relationship between confidence and similarity ratings at the single-participant level. It could be argued that similarity ratings are in fact a better, truer measure of performance because they reflect how similarly two faces are perceived by a person in an ecologically valid setting,” admit the researchers. The group suggest that this is surprising, they argue, because it sets facial movements apart from other body movements. They speculate that this distinction might be related to the lack of concurrent visual information during social interactions. Data and code are available from: https://gitlab.com/elisa.filevich/cistonetal_metacognitionoffacialexpressions.
Brain activity in the supramarginal gyrus is a key part of the sociocognitive process of deception.
S. Meier and colleagues (2021) report that the motivation for researching the complex behavior of deception exists to identify mechanisms of sociocognitive functioning. Deception occurs at various levels of society even becoming apparent in current politics. An understanding of the neurobiological aspects of deception has implications for subsequent theory of mind and social cognition research. Results from the activation likelihood estimation identified seven brain regions significantly activated during deception. Across the varying studies, they found significant activation in the insula, superior and medial frontal gyri, and supramarginal gyrus.
There were 3400 papers involved in the study.
Neuroscientists have shown that V1 neurons in the substantia nigra respond differently to different types of stimulus, such as natural scenes and white Gaussian noise.
In ‘How Stimulus Statistics Affect The Receptive Fields of V1 Cells’, a research team from the National Vision Research Institute led by A. Almasi (2021) reported that the understanding of sensory coding in the visual system is largely based on the stimulus-response characterization of neurons. The researchers studied the spatial structure of V1 receptive fields in anaesthetized cats using both White Gaussian Noise and Natural Scenes as stimuli. They found that the RF structure uncovered using NS resulted in more filters than did WGN. They studied the differences in the RF characteristics of cells in cat V1 when stimulated with white Gaussian noise (WGN) and natural scenes (NS). The authors believe it is attributed to the wider range of feature-contrasts generated by RF filters when using NS stimuli.
It should be noted that this analysis was done for cells whose RFs were uncovered using both WGN and NS, but had more filters using NS. After matching cells’ spike counts, they performed the statistical significance test to determine the number of spatial filters within the RF. The findings appear to reinforce previous work in this area: “R.R. Shapley argued that adaptation to scene statistics is equivalent to contrast adaptation. This is pertinent to the present study in which differences between the statistics of WGN and NS led to differences in the contrasts of the corresponding features in each cell’s RF filters,” Almasi said.
A drug that blocks the activity of the sirtuin pathway has been shown to improve the swimming behaviour of a zebrafish model of Machado-Joseph disease.
M. Watchon et al. (2021) report that Machado-Joseph disease, known as spinocerebellar ataxia-3, is a fatal neurodegenerative disease characterized by clinical symptoms including ataxia, dystonia, rigidity, muscle atrophy and visual and speech disorder. Co-treating MJD zebrafish with sodium valproate and EX527 prevented the increased LC3II/I ratio produced by SV treatment alone. The authors found that this beneficial effect of SV treatment was occurring through activity of the sirtuin pathway. SV treatment induces increased activity of the sirtuin pathway, which has protective effects for MJD zebrafish motor function. Resveratrol treatment produced a significant increase in levels of SIRT1 protein compared to those receiving vehicle control.
However, “Valproic acid had limited benefit for a mouse model of MJD. One trial has reported that treating MJD patients with SV is safe and efficacious. Testing of SV treatment for Huntington's disease has found a dose dependent alleviation of hyperkinesia,” acknowledge the investigators.
A model of the brain's ability to track representational drift aims to explain why the brain is able to self-heal.
In ‘Self-Healing Neural Codes’, M. Rule and T. O'Leary (2021) noted that the cellular and molecular components of the brain change over time. The authors describe a novel form of homeostatic plasticity that allows consolidated representations to interoperate with unstable neural populations. Stable neural codes should homeostatically preserve internal models. Neurons can exploit this to improve their robustness to drift. Even if learning has ceased, these connections remain. The team explore how neural networks could track drift in sensorimotor representations.
The researchers’ results potentially substantiate what was previously known about this field: “This constraint preserves the functional relationships between neural populations. The approach developed here shares some similarities with approaches to attenuate forgetting using replay during sleep, or the equivalent in artificial networks,” Rule suggested.
Neuronavigation software can be used to localise the location of electrodes in a simultaneous EEG-fMRI data set, and suggest that researchers check the regions underlying their neuroimaging results.
C. Scrivener and A. Reader (2021) reported in ‘Variability of EEG electrode positions and their underlying brain regions’ that Electrode C1 was mapped to the left primary motor cortex, whereas C2 was closer to right pre-motor cortex. The authors found variance in electrode placement that was comparable with previous studies, with the largest deviations in the z dimension and in occipital and parietal electrodes. Electrode positions were marked by placing targets onto the centre of the gel artifacts, orthogonal to the skin. The group found a grand standard deviation of 3.94 mm in x, 5.55 mm in y, and 7.17 mm in z. The five electrodes with the smallest overall deviation in xyz were mostly in frontal and central locations (F5, F7, FC5, FCz, FT7).
There were 20 participants included in the study. The researchers’ results claim to challenge prior work in this subject: “frontal electrodes had the smallest deviation across subjects, in co-ordinates both at the scalp and projected onto the brain. However, we did not identify any greater variation specifically in electrodes around the edge of the electrode cap.” Data and code are available from: https://bioimagesuiteweb.github.io/webapp.
Anxiety alters the brain's ability to encode predictions and update beliefs in a probabilistic reward-learning task.
Thomas Hein and Maria Ruiz (2021) described how state anxiety alters the neural oscillatory correlates of predictions and prediction errors during reward learning. Affective states closely interact with decision making. This study investigated how anxiety states modulate the oscillatory correlates of low-level reward predictions and prediction errors during learning in a volatile environment. They tested this by re-analysing data from the previous study, which investigated Bayesian Predictive Coding in state anxiety. The results extend computational work on maladaptive learning in anxiety. State anxiety significantly increased the percentage of errors made during reward learning when compared to the control group. In parallel to the cardiovascular and behavioural changes induced by the anxiety manipulation, by modelling decisions with the HGF, they found that state anxiety impaired learning.
The study involved 42 participants. The researchers’ conclusions may substantiate prior work in this field: “HRV and high-frequency HRV measures were derived from the R-peaks extracted from the EKG signal recorded throughout the experimental sessions. In line with prior research, our previous study showed reduced HF-HRV and reduced HRV in state anxious participants,” Hein suggested. They suggest that future studies may benefit from a stimulus-locked approach when assessing the oscillatory correlates of predictions. Data and code to reproduce the analyses can be found at: https://osf.io/b4qkp.
Analysis of brain activity across five domains of creative production: verbalizing, music, movement, writing, and drawing.
In ‘The neural basis of creative production’, S. Brown and E. Kim (2021) noted that the term ‘creativity’ is both easy to define and difficult to apply. The involvement of sensorimotor areas in creative production has been underappreciated. These areas are interconnected via the frontal aslant tract, suggesting that this tract might be a useful target in white-matter analyses of creativity. In its most universal definition, creativity refers to the novelty of an idea or product, including its originality and level of surprise. Brown and Kim propose distinguishing creation, as a categorical process distinct from replication and imitation, from the graded continua of creativeness.
There were 16 AUT studies included in the study. Discussing potential shortcomings, “Creative processing is an enhancement of brain areas mediating non-creative or less creative processing. Future work on the neuroscience of creativity needs to move beyond brainstorming tasks to look at multi-phase tasks that involve elaboration and exploration in addition to generation,” they concede. They propose that long-term creativity will be an important frontier for future work on the neuroscience of creativity. The pre-SMA and frontal operculum were present in the generation phase, whereas the dorsal part of BA 44 was present in revision, but not generation.
Using functional magnetic resonance imaging (fMRI) to map the whole brain has the potential to shed light on the long-term effects of cannabis use on the brain.
J. Ramaekers et al. (2021) reported on spatial patterns in functional brain connectomes reflect acute and chronic cannabis use. The endocannabinoid system has been implicated to play a modulatory role in cognition and motor function, neuroprotection, nociception, synaptic plasticity and inflammation. The present study aimed to determine the effects of acute and chronic cannabis use on the whole brain connectome. Acute cannabis intoxication produced a select, spatial pattern in functional connectivity that was strongly associated with the feeling of subjective high in both user groups.
The research involved 14 users.
Neuroimaging data from the developing Human Connectome Project (dHCP) has been used to assess the developing brain fingerprint of pre-term babies.
Judit Ciarrusta and colleagues (2021) described the developing brain structural and functional connectome fingerprint. The Human Connectome Project is an observational, cross-sectional Open Science programme approved by the UK National Research Ethics Authority. The brain structural connectome fingerprint is already present in the perinatal period. It is relatively stable and individually unique at this stage of development. Functional connectivity is either too dynamic or immature to provide strong identification features. Future studies should investigate regional differences throughout development.
The research involved 63 subjects. Discussing potential improvements, “The main limitation of this study is that to assess the connectome fingerprint across weeks in the perinatal period ex-utero there is no other option but to investigate a preterm born population. To what extent the fingerprint is affected by preterm birth will have to be investigated in older infant cohorts or using foetal MRI,” they admit. Ciarrusta and colleagues propose that study used anatomical parcellations to characterize functional nodes. Future studies should investigate whether diverse functional parcellation methods mimic or differ from the findings reported in this study. Data and code are available from: https://data.developingconnectome.org.
Negative and positive valence systems (NVS and PVS) are key brain regions involved in decision-making in the immediate aftermath of trauma.
In ‘Differential Roles of Positive and Negative Valence Systems in the Development of Post-Traumatic Stress Psychopathology’, Z. Ben-Zion et al. (2021) noted that the concept of negative and positive valence systems (NVS and PVS, respectively) originated in psychology over a century ago, yet more recently was incorporated into the field of clinical neuroscience. To find the neural indicators of the PVS and NVS associated with Post-Traumatic Stress Disorder symptom severity shortly after exposure, partial correlations were computed between neural indices, and PTSD symptom severity. The authors found that stronger amygdala-lateral orbitofrontal cortex connectivity in the contrast of punishments vs. rewards was associated with more severe symptoms. Traumatic stress might hinder accurate valence estimations, as it increases vigilance and drains cognitive resources. Transition into reward-driven behavior over time, despite the presence of a heightened threat, is necessary for promoting stress resilience.
132 participants were included in the research. The team suggest that future studies should examine the neural response using a network perspective or a data-driven whole-brain approach.