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

Neuroimaging research requires purpose-built analysis software, which is challenging to install and may produce different results across computing environments. The community-oriented, open-source Neurodesk platform ( https://www.neurodesk.org/ ) harnesses a comprehensive and growing suite of neuroimaging software containers. Neurodesk includes a browser-accessible virtual desktop, command-line interface and computational notebook compatibility, allowing for accessible, flexible, portable and fully reproducible neuroimaging analysis on personal workstations, high-performance computers and the cloud.

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 pediatric magnetoencephalography 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-min 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. At the sensor level, the older children showed significantly larger MMF amplitudes relative to the younger children. At the source level, the older children showed a significantly larger MMF amplitude in the right inferior frontal gyrus relative to the younger children, P < 0.05. No differences were found in 2 other key regions (right primary auditory cortex and right superior temporal gyrus) thought to be involved in mismatch generation. 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.

Natural vs forced language switching: Free selection and consistent language use eliminate significant performance costs and cognitive demands in the brain.

Bilinguals are known to switch language spontaneously in everyday conversations, even if there are no external requirements to do so. However, in the laboratory setting, language control is often investigated using forced switching tasks, which result in significant performance costs. The present study assessed whether switching would be less costly when performed in a more natural fashion, and what factors might account for this. Mandarin-English bilinguals engaged in language switching under three different contexts with varied task demands. We examined two factors which may be characteristic of natural switching: (i) freedom of language selection; (ii) consistency of language used to name each item. Participants' brain activities were recorded using magnetoencephalography (MEG), along with behavioural measures of reaction speed and accuracy. The natural context (with both free selection and consistent language use for each item) produced better performance overall, showing reduced mixing cost and no significant switch cost. The neural effect of language mixing was also reversed in this context, suggesting that freely mixing two languages was easier than staying in a single language. Further, while switching in the forced context elicited increased brain activity in the right inferior frontal gyrus, this switch effect disappeared when the language used to name each item was consistent. Together, these findings demonstrate that the two factors above conjointly contribute to eliminating significant performance costs and cognitive demands associated with language switching and mixing. Such evidence aligns with lexical selection models which do not assume bilingual production to be inherently effortful.

Comprehensive behavioural intervention for tics: a neurophysiological intervention.

BACKGROUND: Gilles de la Tourette Syndrome (GTS) is a childhood-onset neuropsychiatric disorder characterised by motor and vocal tics. While Comprehensive Behavioural Intervention for Tics (CBIT) is an effective, non-pharmacological treatment for patients with GTS, the underlying neurophysiological basis of this intervention has not been investigated. METHODS: To investigate the clinical effectiveness of CBIT in reducing tic severity in young people with GTS and explore neurophysiological mechanisms associated with clinical change. RESULTS: There was a significant overall improvement in tic severity of large effect size. The Cortical Silent Period (CSP) to motor evoked potential (MEP) ratio (CSP/MEP ratio) increased after the intervention with a small effect size. Other neurophysiological measures of inhibition were not significantly related to the change in tic severity. CONCLUSIONS: Alongside significant clinical improvements, these results suggest a role for motor cortical Gamma-aminobutyric acid (GABA)-ergic inhibitory circuitry in the neurophysiological changes underlying CBIT treatment. These findings need to be replicated in larger studies using control samples.

A Guided Tutorial on Modelling Human Event-Related Potentials with Recurrent Neural Networks.

In cognitive neuroscience research, computational models of event-related potentials (ERP) can provide a means of developing explanatory hypotheses for the observed waveforms. However, researchers trained in cognitive neurosciences may face technical challenges in implementing these models. This paper provides a tutorial on developing recurrent neural network (RNN) models of ERP waveforms in order to facilitate broader use of computational models in ERP research. To exemplify the RNN model usage, the P3 component evoked by target and non-target visual events, measured at channel Pz, is examined. Input representations of experimental events and corresponding ERP labels are used to optimize the RNN in a supervised learning paradigm. Linking one input representation with multiple ERP waveform labels, then optimizing the RNN to minimize mean-squared-error loss, causes the RNN output to approximate the grand-average ERP waveform. Behavior of the RNN can then be evaluated as a model of the computational principles underlying ERP generation. Aside from fitting such a model, the current tutorial will also demonstrate how to classify hidden units of the RNN by their temporal responses and characterize them using principal component analysis. Statistical hypothesis testing can also be applied to these data. This paper focuses on presenting the modelling approach and subsequent analysis of model outputs in a how-to format, using publicly available data and shared code. While relatively less emphasis is placed on specific interpretations of P3 response generation, the results initiate some interesting discussion points.

Increased segregation of functional networks in developing brains.

A growing literature conceptualises typical brain development from a network perspective. However, largely due to technical and methodological challenges inherent in paediatric functional neuroimaging, there remains an important gap in our knowledge regarding the typical development of functional brain networks in "preschool" childhood (i.e., children younger than 6 years of age). In this study, we recorded brain oscillatory activity using age-appropriate magnetoencephalography in 24 children, including 14 preschool children aged from 4 to 6 years and 10 school children aged from 7 to 12 years. We compared the topology of the resting-state brain networks in these children, estimated using minimum spanning tree (MST) constructed from phase synchrony between beamformer-reconstructed time-series, with that of 24 adults. Our results show that during childhood the MST topology shifts from a star-like (centralised) toward a more line-like (de-centralised) configuration, indicating the functional brain networks become increasingly segregated. In addition, the increasing global network segregation is frequency-independent and accompanied by decreases in centrality (or connectedness) of cortical regions with age, especially in areas of the default mode network. We propose a heuristic MST model of "network space", which posits a clear developmental trajectory for the emergence of complex brain networks. Our results not only revealed topological reorganisation of functional networks across multiple temporal and spatial scales in childhood, but also fill a gap in the literature regarding neurophysiological mechanisms of functional brain maturation during the preschool years of childhood.

Adaptive Motor Imagery: A Multimodal Study of Immobilization-Induced Brain Plasticity.

The consequences of losing the ability to move a limb are traumatic. One approach that examines the impact of pathological limb nonuse on the brain involves temporary immobilization of a healthy limb. Here, we investigated immobilization-induced plasticity in the motor imagery (MI) circuitry during hand immobilization. We assessed these changes with a multimodal paradigm, using functional magnetic resonance imaging (fMRI) to measure neural activation, magnetoencephalography (MEG) to track neuronal oscillatory dynamics, and transcranial magnetic stimulation (TMS) to assess corticospinal excitability. fMRI results show a significant decrease in neural activation for MI of the constrained hand, localized to sensorimotor areas contralateral to the immobilized hand. MEG results show a significant decrease in beta desynchronization and faster resynchronization in sensorimotor areas contralateral to the immobilized hand. TMS results show a significant increase in resting motor threshold in motor cortex contralateral to the constrained hand, suggesting a decrease in corticospinal excitability in the projections to the constrained hand. These results demonstrate a direct and rapid effect of immobilization on MI processes of the constrained hand, suggesting that limb nonuse may not only affect motor execution, as evidenced by previous studies, but also MI. These findings have important implications for the effectiveness of therapeutic approaches that use MI as a rehabilitation tool to ameliorate the negative effects of limb nonuse.

Abnormal time course of low beta modulation in non-fluent preschool children: A magnetoencephalographic study of rhythm tracking.

Stuttering is a disorder of speech affecting millions of people around the world. Whilst the exact aetiology of stuttering remains unknown, it has been hypothesised that it is a disorder of the neural mechanisms that support speech timing. In this article, we used magnetoencephalography (MEG) to examine activity from auditory regions of the brain in stuttering and non-stuttering children aged 3-9years. For typically developing children, we found that MEG oscillations in the beta band responded to rhythmic sounds with a peak near the time of stimulus onset. In contrast, stuttering children showed an opposite phase of beta band envelope, with a trough of activity at stimulus onset. These results suggest that stuttering may result from abnormalities in predictive brain responses which are reflected in abnormal entrainment of the beta band envelope to rhythmic sounds.

Vocal response inhibition is enhanced by anodal tDCS over the right prefrontal cortex.

Stopping outright (reactive inhibition) and slowing down (proactive inhibition) are types of response inhibition which have mainly been investigated in the manual effector system. This study compared reactive inhibition across manual and vocal effector systems, examined the effects of excitatory anodal transcranial direct current stimulation (anodal tDCS) over the right prefrontal cortex (right-PFC) and looked at the relationship between reactive and proactive inhibition. We hypothesised (1) that vocal reactive inhibition would be less effective than manual reactive inhibition as evidenced by longer stop signal reaction times; (2) that anodal tDCS would enhance both vocal and manual reactive inhibitions and (3) that proactive and reactive inhibitions would be positively related. We tested 14 participants over two sessions (one session with anodal tDCS and one session with sham stimulation) and applied stimulation protocol in the middle of the session, i.e. only during the second of three phases. We used a stop signal task across two stop conditions: relevant and irrelevant stop conditions in which stopping was required or ignored, respectively. We found that reactive inhibition was faster during and immediately after anodal tDCS relative to sham. We also found that greater level of proactive inhibition enhanced reactive inhibition (indexed by shorter stop signal reaction times). These results support the hypothesis that the right-PFC is part of a core network for reactive inhibition and supports previous contention that proactive inhibition is possibly modulated via preactivating the reactive inhibition network.

Event-related fields evoked by vocal response inhibition: a comparison of younger and older adults.

The current study examined event-related fields (ERFs) evoked by vocal response inhibition in a stimulus-selective stop-signal task. We compared inhibition-related ERFs across a younger and an older group of adults. Behavioural results revealed that stop-signal reaction times (RTs), go-RTs, ignore-stop RTs and failed stop RTs were longer in the older, relative to the younger group by 38, 123, 149 and 116 ms, respectively. The amplitude of the ERF M2 peak (approximately 200 ms after the stop signal) evoked on successful stop trials was larger compared to that evoked on both failed stop and ignore-stop trials. The M4 peak (approximately 450 ms after stop signal) was of larger amplitude in both successful and failed stops compared to ignore-stop trials. In the older group, the M2, M3 and M4 peaks were smaller in amplitude and peaked later in time (by 24, 50 and 76 ms, respectively). We demonstrate that vocal response inhibition-related ERFs exhibit a similar temporal evolution to those previously described for manual response inhibition: an early peak at 200 ms (i.e. M2) that differentiates successful from failed stopping, and a later peak (i.e. M4) that is consistent with a neural marker of response checking and error processing. Across groups, our data support a more general decline of stimulus processing speed with age.

Development of effective connectivity in the core network for face perception.

This study measured effective connectivity within the core face network in young children using a paediatric magnetoencephalograph (MEG). Dynamic casual modeling (DCM) of brain responses was performed in a group of adults (N = 14) and a group of young children aged from 3 to 6 years (N = 15). Three candidate DCM models were tested, and the fits of the MEG data to the three models were compared at both individual and group levels. The results show that the connectivity structure of the core face network differs significantly between adults and children. Further, the relative strengths of face network connections were differentially modulated by experimental conditions in the two groups. These results support the interpretation that the core face network undergoes significant structural configuration and functional specialization between four years of age and adulthood.

A conceptual lemon: theta burst stimulation to the left anterior temporal lobe untangles object representation and its canonical color.

Object recognition benefits greatly from our knowledge of typical color (e.g., a lemon is usually yellow). Most research on object color knowledge focuses on whether both knowledge and perception of object color recruit the well-established neural substrates of color vision (the V4 complex). Compared with the intensive investigation of the V4 complex, we know little about where and how neural mechanisms beyond V4 contribute to color knowledge. The anterior temporal lobe (ATL) is thought to act as a "hub" that supports semantic memory by integrating different modality-specific contents into a meaningful entity at a supramodal conceptual level, making it a good candidate zone for mediating the mappings between object attributes. Here, we explore whether the ATL is critical for integrating typical color with other object attributes (object shape and name), akin to its role in combining nonperceptual semantic representations. In separate experimental sessions, we applied TMS to disrupt neural processing in the left ATL and a control site (the occipital pole). Participants performed an object naming task that probes color knowledge and elicits a reliable color congruency effect as well as a control quantity naming task that also elicits a cognitive congruency effect but involves no conceptual integration. Critically, ATL stimulation eliminated the otherwise robust color congruency effect but had no impact on the numerical congruency effect, indicating a selective disruption of object color knowledge. Neither color nor numerical congruency effects were affected by stimulation at the control occipital site, ruling out nonspecific effects of cortical stimulation. Our findings suggest that the ATL is involved in the representation of object concepts that include their canonical colors.

Neuromagnetic brain activity associated with anticipatory postural adjustments for bimanual load lifting.

During bimanual load lifting, the brain must anticipate the effects of unloading upon the load-bearing arm. Little is currently known about the neural networks that coordinate these anticipatory postural adjustments. We measured neuromagnetic brain activity with whole-head magnetoencephalography while participants performed a bimanual load-lifting task. Anticipatory adjustments were associated with reduction in biceps brachii muscle activity of the load-bearing arm and pre-movement desynchronization of the cortical beta rhythm. Beamforming analyses localized anticipatory brain activity to the precentral gyrus, basal ganglia, supplementary motor area, and thalamus, contralateral to the load-bearing arm. To our knowledge this is the first human neuroimaging study to directly investigate anticipatory postural adjustments and to explicitly partition the anticipatory and volitional aspects of brain activity in bimanual load lifting. These data contribute to our understanding of the neural systems supporting anticipatory postural adjustments in healthy adults.

Filling your time: Auditory flutter alters perceived duration via stimulus-locked responses.

In the current article, we examined the flutter-duration illusion; the extension of perceived duration when an interval is filled with auditory flutter. Participants reproduced flutter-filled and empty durations while electrophysiological activity was recorded. As expected, participants over-produced durations when they were filled with auditory flutter rather than unfilled. Using multivariate pattern analysis, we found several differentiating patterns of neural activity while participants listened to either flutter-filled or empty intervals. However, in subsequent single trial analysis, only two of these clusters predicted perceived duration in the flutter condition; one occurring in line with the second click of the flutter, and one in line with the fourth click. We relate this finding to the N1P2 component and P3a component to timing initiation and arousal, respectively.

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

Objective. To use a recurrent neural network (RNN) to reconstruct neural activity responsible for generating noninvasively measured electromagnetic signals.Approach. Output 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 minimise 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 regularisation 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 beamformers, minimum-norm estimate, and mixed-norm estimate source reconstruction methods.Main results. The proposed RNN method had higher output signal-to-noise ratios and comparable correlation and error between estimated and simulated sources. Reconstructed MEG signals were also equal or superior to the other methods regarding 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 biophysically plausible and consistent with expectations from the literature.Significance. This 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 localised neural activity.

Theta tACS impairs episodic memory more than tDCS.

Episodic memory deficits are a common consequence of aging and are associated with a number of neurodegenerative disorders (e.g., Alzheimer's disease). Given the importance of episodic memory, a great deal of research has investigated how we can improve memory performance. Transcranial electrical stimulation (TES) represents a promising tool for memory enhancement but the optimal stimulation parameters that reliably boost memory are yet to be determined. In our double-blind, randomised, sham-controlled study, 42 healthy adults (36 females; 23.3 ± 7.7 years of age) received anodal transcranial direct current stimulation (tDCS), theta transcranial alternating current stimulation (tACS) and sham stimulation during a list-learning task, over three separate sessions. Stimulation was applied over the left temporal lobe, as encoding and recall of information is typically associated with mesial temporal lobe structures (e.g., the hippocampus and entorhinal cortex). We measured word recall within each stimulation session, as well as the average number of intrusion and repetition errors. In terms of word recall, participants recalled fewer words during tDCS and tACS, compared to sham stimulation, and significantly fewer words recalled during tACS compared with tDCS. Significantly more memory errors were also made during tACS compared with sham stimulation. Overall, our findings suggest that TES has a deleterious effect on memory processes when applied to the left temporal lobe.

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

Objective.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 aimed to develop a computational model of visual ERP generation to study this phenomenon which consisted of a three-dimensional convolutional neural network (CNN) connected to a recurrent neural network (RNN).Approach.The CNN provided image representation learning, complimenting sequence learning of the RNN for modeling visually-evoked potentials. We used open-access data from ERP Compendium of Open Resources and Experiments (40 subjects) to develop the model, generated synthetic images for simulating experiments with a generative adversarial network, then collected additional data (16 subjects) 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. By filtering and pooling over spatial dimensions, 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.Main results.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.Significance.The approach developed in this work is potentially of significant value for visual neuroscience research, where it may be adapted for multiple contexts to study computational relationships between visual stimuli and evoked neural activity.

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.

Self-initiation and temporal cueing of monaural tones reduce the auditory N1 and P2.

Event-related potentials (ERPs) to tones that are self-initiated are reduced in their magnitude in comparison with ERPs to tones that are externally generated. This phenomenon has been taken as evidence for an efference copy of the motor command acting to suppress the sensory response. However, self-initiation provides a strong temporal cue for the stimulus which might also contribute to the ERP suppression for self-initiated tones. The current experiment sought to investigate the suppression of monaural tones by temporal cueing and also whether the addition of self-initiation enhanced this suppression. Lastly, the experiment sought to investigate the lateralisation of the ERP suppression via presenting these monaural tones to each ear respectively. We examined source waveforms extracted from the lateralised auditory cortices and measured the modulation of the N1 and P2 components by cueing and self-initiation. Self-initiation significantly reduced the amplitude of the N1 component. Temporal cueing without self-initiation significantly reduced the P2 component. There were no significant differences in the amplitude of either the N1 or the P2 between self-initiation and temporal cuing. There was a significant lateralisation effect on the N1-it being significantly larger contralateral to the ear of stimulation. There was no interaction between lateralisation and side of the temporal cue or side of self-initiation suggesting that the effects of self-initiation and temporal cuing are equal bilaterally. We conclude that a significant proportion of ERP suppression by self-initiation is a result of inherent temporal cueing.