Biology subjects

Sowman, P. F.

Publications and source records attributed to Sowman, P. F..

14 recordsLinked to original sources

bioRxiv2026

Disruption of prefrontal cortex accelerates sensory anchoring but dissociates online recognition from implicit consolidation of sound sequences

The auditory brain excels at extracting predictive regularities from sound, but the neural and temporal mechanisms mediating long-term learning of these regularities remain poorly understood. Here, using a perceptual anchoring paradigm, we investigate whether the dorsolateral prefrontal cortex (dlPFC) is causally required to transform transient regularities into stable internal templates (anchors). We characterized the behavioural dynamics of anchoring to predicable acoustic events using sensitivity (d) and decision bias (c) metrics, alongside growth curve analysis (GCA) to model trial-by-trial learning. Listeners exhibited a robust anchoring effect, showing significantly higher sensitivity in decision strategy for recurring acoustic regularities compared to novel ones. This implicit anchoring was driven selectively by specific features in regular sequences--the local repetition of an acoustic motif--whereas explicit memory indiscriminately encoded the entire acoustic episode. To test for causality, we applied inhibitory continuous theta-burst stimulation (cTBS) to the dlPFC. Crucially, not only did this not impair anchoring, but significantly accelerated the rate of anchoring for predictable patterns, suggesting the dlPFC may act as a source of top-down interference in implicit learning. In contrast to predictable sequences, inhibition of dlPFC disrupted the initial precision of online comparisons for novel sequences, though insulating against the rapid, interference-driven reduction in performance or novel sequences in observed in listeners subjected to a sham cTBS procedure. These findings reveal a functional dissociation. Whilst dlPFC provides an essential buffer for novel sensory comparisons, long-term consolidation of regularities in acoustic sequences relies on a robust implicit pathway independent of prefrontal control.

neuroscience↗

bioRxiv2024

Low-frequency hearing thresholds improve as high-frequency hearing sensitivity deteriorates between young adulthood and middle age in normally hearing people

Hearing sensitivity changes throughout a persons lifetime. This work aimed to describe changes in pure-tone audiometric (PTA) thresholds that occur in the transition from young adulthood to middle age in 121 adults with normal or nearly normal hearing. Results showed that older people had worse high-frequency (4000-8000 Hz) thresholds and better low-frequency (125-500 Hz) thresholds than younger individuals, suggesting that hearing sensitivity in the low-frequency range may improve with age. The improvement of low-frequency thresholds may be part of a central compensation for age-related deterioration of high-frequency hearing sensitivity. Further investigation of age-related changes in low-frequency hearing sensitivity is needed to confirm our findings.

physiology↗

bioRxiv2024

Localized estimation of event-related neural source activity from simultaneous MEG-EEG with a recurrent neural network

Estimating intracranial current sources underlying the electromagnetic signals observed from extracranial sensors is a perennial challenge in non-invasive neuroimaging. Established solutions to this inverse problem treat time samples independently without considering the temporal dynamics of event-related brain processes. This paper describes current source estimation from simultaneously recorded magneto- and electro-encephalography (MEEG) using a recurrent neural network (RNN) that learns sequential relationships from neural data. The RNN was trained in two phases: (1) pre-training and (2) transfer learning with L1 regularization applied to the source estimation layer. Performance of using scaled labels derived from MEEG, magnetoencephalography (MEG), or electroencephalography (EEG) were compared, as were results from volumetric source space with free dipole orientation and surface source space with fixed dipole orientation. Exact low-resolution electromagnetic tomography (eLORETA) and mixed-norm L1/L2 (MxNE) source estimation methods were also applied to these data for comparison with the RNN method. The RNN approach outperformed other methods in terms of output signal-to-noise ratio, correlation and mean-squared error metrics evaluated against ground-truth event-related field (ERF) and event-related potential (ERP) waveforms. Using MEEG labels with fixed-orientation surface sources produced the most consistent estimates. To estimate sources of ERF and ERP waveforms, the RNN generates temporal dynamics within its internal computational units, driven by sequential structure in neural data used as training labels. It thus provides a data-driven model of computational transformations from psychophysiological events into corresponding event-related neural signals, which is unique among MEG and EEG source reconstruction solutions.

neuroscience↗

bioRxiv2024

Estimating Directed Connectivity with OPM-MEG: A Feasibility Study

We explored the directed connectivity of magnetoencephalography data recorded via Optically Pumped Magnetometers (OPM-MEG). Ten healthy adult participants were scanned twice while watching a movie in an OPM-MEG system. Beamformer source localisation was employed to obtain source time series within 62 cortical brain regions. Multivariate transfer entropy (mTE) was used to quantify directed connectivity. Our results suggest that out-degree connectivity (node-to-network) was stronger than in-degree (network-to-node). Out-degree display a more heavy-tailed distribution, indicative of brain hubs, whereas in-degree nodes had a distribution closer to a normal distribution. Although in-degree and out-degree nodes were positively correlated, we observed that out-degree connectivity was stronger in somatomotor and attention networks than in other resting-state networks. However, in- and out-degree were both inversely related to the physical distance between nodes, suggesting that directed connectivity is stronger between short-distanced nodes. As OPM-MEG research is likely to expand substantially in the future, we propose that directed connectivity measures may provide additional insights into the underlying mechanisms of brain communication.

neuroscience↗

bioRxiv2023

Localized estimation of electromagnetic sources underlying event-related fields using recurrent neural networks

ObjectiveTo use a recurrent neural network (RNN) to reconstruct neural activity responsible for generating noninvasively measured electromagnetic signals. ApproachOutput weights of an RNN were fixed as the lead field matrix from volumetric source space computed using the boundary element method with co-registered structural magnetic resonance images and magnetoencephalography (MEG). Initially, the network was trained to minimize mean-squared-error loss between its outputs and MEG signals, causing activations in the penultimate layer to converge towards putative neural source activations. Subsequently, L1 regularization was applied to the final hidden layer, and the model was fine-tuned, causing it to favour more focused activations. Estimated source signals were then obtained from the outputs of the last hidden layer. We developed and validated this approach with simulations before applying it to real MEG data, comparing performance with three existing methods: beamformers, minimum-norm estimate, and dynamical statistical parametric mapping. Main resultsThe proposed method had higher output signal-to-noise ratios than the others and comparable correlation and error between estimated and simulated sources. Reconstructed MEG signals were also equal or superior to the other methods in terms of their similarity to ground-truth. When applied to MEG data recorded during an auditory roving oddball experiment, source signals estimated with the RNN were generally consistent with expectations from the literature and qualitatively smoother and more reasonable-looking than estimates from the other methods. SignificanceThis work builds on recent developments of RNNs for modelling event-related neural responses by incorporating biophysical constraints from the forward model, thus taking a significant step towards greater biological realism and introducing the possibility of exploring how input manipulations may influence localized neural activity.

neuroscience↗

bioRxiv2023

Neural correlates of face perception modeled with a convolutional recurrent neural network

Event-related potential (ERP) sensitivity to faces is predominantly characterized by an N170 peak that has greater amplitude and shorter latency when elicited by human faces than images of other objects. We developed a computational model of visual ERP generation to study this phenomenon which consisted of a convolutional neural network (CNN) connected to a recurrent neural network (RNN). We used open-access data to develop the model, generated synthetic images for simulating experiments, then collected additional data to validate predictions of these simulations. For modeling, visual stimuli presented during ERP experiments were represented as sequences of images (time x pixels). These were provided as inputs to the model. The CNN transformed these inputs into sequences of vectors that were passed to the RNN. The ERP waveforms evoked by visual stimuli were provided to the RNN as labels for supervised learning. The whole model was trained end-to-end using data from the open-access dataset to reproduce ERP waveforms evoked by visual events. Cross-validation model outputs strongly correlated with open-access (r = 0.98) and validation study data (r = 0.78). Open-access and validation study data correlated similarly (r = 0.81). Some aspects of model behavior were consistent with neural recordings while others were not, suggesting promising albeit limited capacity for modeling the neurophysiology of face-sensitive ERP generation.

neuroscience↗

bioRxiv2022

Neurodesk: An accessible, flexible, and portable data analysis environment for reproducible neuroimaging

Neuroimaging data analysis often requires purpose-built software, which can be challenging to install and may produce different results across computing environments. Beyond being a roadblock to neuroscientists, these issues of accessibility and portability can hamper the reproducibility of neuroimaging data analysis pipelines. Here, we introduce the Neurodesk platform, which harnesses software containers to support a comprehensive and growing suite of neuroimaging software (https://www.neurodesk.org/). Neurodesk includes a browser-accessible virtual desktop environment and a command line interface, mediating access to containerized neuroimaging software libraries on various computing platforms, including personal and high-performance computers, cloud computing and Jupyter Notebooks. This community-oriented, open-source platform enables a paradigm shift for neuroimaging data analysis, allowing for accessible, flexible, fully reproducible, and portable data analysis pipelines.

neuroscience↗

bioRxiv2022

Hearing the world differently: Examining predictive coding accounts of autism using MEG

Predictive coding accounts of autism suggest that autistic perception is characterised by divergent precision weighting. The precise nature of this divergence, however, is debated. Here, we sought to disentangle competing predictive coding accounts of autism by testing them at a neural level. To this end, we used paediatric magnetoencephalography to record the auditory evoked fields of 10 young autistic children (M = 6.2 years, range = 4.2- 8.6) and 63 neurotypical children (M = 6.1 years, range = 3.0- 9.8) as they listened to a roving auditory oddball paradigm. For each participant, we subtracted the evoked responses to the standard from the deviant pure tones to calculate the mismatch field MMF: an electrophysiological component that is widely interpreted as a neural signature of predictive coding. We found no significant differences between the two groups MMF amplitudes, p > .05. An exploratory analysis indicated larger MMF amplitudes in most of the autistic children compared to their average-age-matched neurotypical counterparts, p < .05. We interpret these findings as preliminary evidence in support of the inflexibly high prior and sensory precision account, and against the inflexibly low prior-relative-to-sensory precision accounts of autistic perception. HighlightsO_LIWe used paediatric MEG to compare autistic and neurotypical MMFs amplitudes. C_LIO_LIExploratory case-cohort analyses revealed mostly larger MMFs in autistic cases. C_LIO_LILarger MMFs support the notion of precise, inflexible prediction errors in autism. C_LI

neuroscience↗

bioRxiv2022

Investigating Predictive Coding in Younger and Older Children Using MEG and a Multi-Feature Auditory Oddball Paradigm

There is mounting evidence for predictive coding theory from computational, neuroimaging, and psychological research. However there remains a lack of research exploring how predictive brain function develops across childhood. To address this gap, we used paediatric magnetoencephalography (MEG) to record the evoked magnetic fields of 18 younger children (M = 4.1 years) and 19 older children (M = 6.2 years) as they listened to a 12-minute auditory oddball paradigm. For each child, we computed a mismatch field MMF: an electrophysiological component that is widely interpreted as a neural signature of predictive coding. Consistent with our hypotheses, the older children showed significantly larger MMF amplitudes relative to the younger children. Furthermore, the older children showed a significantly larger MMF amplitude in the right inferior frontal gyrus (IFG; 0.312 to 0.33 s) relative to the younger children, p < .05. These findings support the idea that predictive brain function develops during childhood, with increasing involvement of the frontal cortex in response to prediction errors. These findings contribute to a deeper understanding of the brain function underpinning child cognitive development. HighlightsO_LIThis is the first paediatric MEG study to examine the sources underlying the MMF. C_LIO_LIOlder children showed larger MMF amplitudes in the right inferior frontal gyrus. C_LIO_LIResults support the idea that predictive brain function develops during childhood. C_LI

neuroscience↗

bioRxiv2021

Cortical Oscillatory Dysrhythmias in Visual Snow Syndrome: A MEG Study

Visual Snow (VS) refers to the persistent visual experience of static in the whole visual field of both eyes. It is often reported by patients with migraine and co-occurs with conditions like tinnitus and tremor. The underlying pathophysiology of the condition is poorly understood. Previously we hypothesised, that VSS may be characterised by disruptions to rhythmical activity within the visual system1. To test this, data from 18 patients diagnosed with visual snow syndrome (VSS), and 16 matched controls, were acquired using Magnetoencephalography (MEG). Participants were presented with visual grating stimuli, known to elicit decreases in alpha-band (8-13Hz) power and increases in gamma-band power (40-70Hz). Data were mapped to source-space using a beamformer. Across both groups, decreased alpha power and increased gamma power localised to early visual cortex. Data from primary visual cortex (V1) were compared between groups. No differences were found in either alpha or gamma peak frequency or the magnitude of alpha power, p>.05. However, compared with controls, our VSS cohort displayed significantly increased V1 gamma power, p=.035. This new electromagnetic finding concurs with previous fMRI and PET findings suggesting that in VSS, the visual cortex is hyper-excitable. The coupling of alpha-phase to gamma amplitude (i.e., phase-amplitude coupling, PAC) within V1 was also quantified. Compared with controls, the VSS group had significantly reduced alpha-gamma PAC, p<.05, indicating a potential excitation-inhibition imbalance in VSS, as well as a potential disruption to top-down "noise-cancellation" mechanisms. Overall, these results suggest that rhythmical brain activity in primary visual cortex is both hyperexcitable and disorganised in VSS, consistent with visual snow being a condition of thalamocortical dysrhythmia.

neuroscience↗

bioRxiv2020

Analysis methods for measuring fNIRS responses generated by a block-design paradigm

OverviewO_ST_ABSSignificanceC_ST_ABSfNIRS is an increasingly popular tool in auditory research, but the range of analysis procedures employed across studies complicates interpretation of data. AimTo assess the impact of different analysis procedures on the morphology, detection, and lateralization of auditory responses in fNIRS. Specifically, whether averaging or GLM-based analyses generate different experimental conclusions, when applied to a block-protocol design. The impact of parameter selection of GLMs on detecting auditory-evoked responses was also quantified. Approach17 listeners were exposed to three commonly employed auditory stimuli: noise, speech, and silence. A block design was employed, comprising sounds of 5-s duration, and 10-20 s silent intervals. ResultsBoth analysis procedures generated similar response morphologies and amplitude estimates, and both also indicated responses to speech to be significantly greater than to noise and silence. Neither approach indicated a significant effect of brain hemisphere on responses to speech. Methods to correct for systemic hemodynamic responses using short channels improved detection at the individual level. ConclusionsConsistent with theoretical considerations, simulations, and other experimental domains, GLM and averaging analyses generate the same group-level experimental conclusions. We release this dataset publicly for use in future development and optimization of algorithms.

neuroscience↗

bioRxiv2019

Co-Increasing Neuronal Noise and Beta Power in the Developing Brain

Accumulating evidence across species indicates that brain oscillations are superimposed upon an aperiodic 1/f - like power spectrum. Maturational changes in neuronal oscillations have not been assessed in tandem with this underlying aperiodic spectrum. The current study uncovers co-maturation of the aperiodic component alongside the periodic components (oscillations) in spontaneous magnetoencephalography (MEG) data. Beamformer-reconstructed MEG time-series allowed a direct comparison of power in the source domain between 24 children (8.0 {+/-} 2.5 years, 17 males) and 24 adults (40.6 {+/-} 17.4 years, 16 males). Our results suggest that the redistribution of oscillatory power from lower to higher frequencies that is observed in childhood does not hold once the age-related changes in the aperiodic signal are controlled for. When estimating both the periodic and aperiodic components, we found that power increases with age in the beta band only, and that the 1/f signal is flattened in adults compared to children. These results suggest a pattern of co-maturing beta oscillatory power with the aperiodic 1/f signal in typical childhood development.

neuroscience↗

bioRxiv2019

Reduced Auditory Steady State Responses in Autism Spectrum Disorder

BackgroundAuditory steady state responses (ASSRs) are elicited by clicktrains or amplitude-modulated tones, which entrain auditory cortex at their specific modulation rate. Previous research has reported reductions in ASSRs at 40Hz for autism spectrum disorder (ASD) participants and first-degree relatives of people diagnosed with ASD [1,2]. MethodsUsing a 1.5s-long auditory clicktrain stimulus, designed to elicit an ASSR at 40Hz, this study attempted to replicate and extend these findings. Magnetencephalography (MEG) data were collected from 18 adolescent ASD participants and 18 typically developing controls. ResultsThe ASSR localised to bilateral primary auditory regions. Regions of interest were thus defined in left and right primary auditory cortex (A1). While the transient gamma-band response (tGBR) from 0-0.1s following presentation of the clicktrain stimulus was not different between groups, for either left or right A1, the ASD group had reduced oscillatory power at 40Hz from 0.5 to 1.5s post-stimulus onset, for both left and right A1. Additionally, the ASD group had reduced inter-trial coherence (phase consistency over trials) at 40Hz from 0.64-0.82s for right A1 and 1.04-1.22s for left A1. LimitationsIn this study, we did not conduct a clinical autism assessment (e.g. the ADOS), and therefore it remains unclear whether ASSR power and/or ITC are associated with the clinical symptoms of ASD. ConclusionOverall, our results support a specific reduction in ASSR oscillatory power and inter-trial coherence in ASD, rather than a generalised deficit in gamma-band responses. We argue that this could reflect a developmentally relevant reduction in non-linear neural processing.

neuroscience↗

bioRxiv2018

Topological segregation of functional networks increases in developing brains

A growing literature conceptualises human brain development from a network perspective, but it remains unknown how functional brain networks are refined during the preschool years. The extant literature diverges in its characterisation of functional network development, with little agreement between haemodynamic- and electrophysiology-based measures. In children aged from 4 to 12 years, as well as adults, age appropriate magnetoencephalography was used to estimate unbiased network topology, using minimum spanning tree (MST) constructed from phase synchrony between beamformer-reconstructed time-series. During childhood, network topology becomes increasingly segregated, while cortical regions decrease in centrality. We propose a heuristic MST model, in which a clear developmental trajectory for the emergence of complex brain networks is delineated. Our results resolve topological reorganisation of functional networks across temporal and special scales in youth and fill a gap in the literature regarding neurophysiological mechanisms of functional brain maturation during the preschool years.

neuroscience↗