Delineating between-subject heterogeneity in alpha networks with Spatio-Spectral Eigenmodes

•

A data-driven modal decomposition describes oscillations by their resonant frequency, damping time and network structure.

•

We show that the full multivariate transfer function can be rewritten as a linear superposition of these modes.

•

These modal coordinates factorise oscillatory systems without pre-specification of frequency bands or regions of interest.

•

Using these modes, we find a spatial gradient in alpha peak frequency between Occipital and Parietal cortex .

•

This gradient is highly variable between participants, showing shifts in spatial structure and peak frequency.

Between subject variability in the spatial and spectral structure of oscillatory networks can be highly informative but poses a considerable analytic challenge. Here, we describe a data-driven modal decomposition of a multivariate autoregressive model that simultaneously identifies oscillations by their peak frequency, damping time and network structure. We use this decomposition to define a set of Spatio-Spectral Eigenmodes (SSEs) providing a parsimonious description of oscillatory networks. We show that the multivariate system transfer function can be rewritten in these modal coordinates, and that the full transfer function is a linear superposition of all modes in the decomposition. The modal transfer function is a linear summation and therefore allows for single oscillatory signals to be isolated and analysed in terms of their spectral content, spatial distribution and network structure. We validate the method on simulated data and explore the structure of whole brain oscillatory networks in eyes-open resting state MEG data from the Human Connectome Project. We are able to show a wide between participant variability in peak frequency and network structure of alpha oscillations and show a distinction between occipital ’high-frequency alpha’ and parietal ’low-frequency alpha’. The frequency difference between occipital and parietal alpha components is present within individual participants but is partially masked by larger between subject variability; a 10Hz oscillation may represent the high-frequency occipital component in one participant and the low-frequency parietal component in another. This rich characterisation of individual neural phenotypes has the potential to enhance analyses into the relationship between neural dynamics and a person’s behavioural, cognitive or clinical state.

A quick and reliable estimate of extended high-frequency hearing

OBJECTIVE

To encourage researchers to perform high-frequency threshold estimation using a technique outlined by Rieke and colleagues, described as fixed-level frequency threshold estimation. Their method used a Bekesy-style roving tone to estimate the highest audible frequency of a listener. The tone was fixed in its intensity (SPL) and changed in frequency as the participant indicated whether they could perceive the tone, or not. This was developed specifically for ototoxicity monitoring in the extended high-frequency region. Rieke and colleagues established that this approach to measuring hearing thresholds is both fast and reliable.

DESIGN

The current article extends this approach to using a simple PC-soundcard-transducer setup and the method of limits to rapidly establish the highest audible frequency of a listener.Study sample: 24 listeners performed standard and fixed-level audiometry in the extended high-frequency range.

RESULTS

The method described is rapid and reliable and a single summary metric is obtained for each listener.

CONCLUSIONS

The advantage of the described approach over standard pure-tone audiometry in the extended high-frequency range is the time taken, the ability to avoid missing data points and the risk of distortions or electrical noise when close to maximal system output.

Charting the effects of TMS with fMRI: Modulation of cortical recruitment within the distributed network supporting semantic control

Semantic memory comprises our knowledge of the meanings of words and objects but only some of this knowledge is relevant at any given time. Thus, semantic control processes are needed to focus retrieval on relevant information. Research on the neural basis of semantic control has strongly implicated left inferior frontal gyrus (LIFG) but recent work suggests that a wider network supports semantic control, including left posterior middle temporal gyrus (pMTG), right inferior frontal gyrus (RIFG) and pre-supplementary motor area (pre-SMA). In the current study, we used repetitive transcranial magnetic stimulation (1 Hz offline TMS) over LIFG, immediately followed by fMRI, to examine modulation of the semantic network. We compared the effect of stimulation on judgements about strongly-associated words (dog-bone) and weaker associations (dog-beach), since previous studies have found that dominant links can be recovered largely automatically with little engagement of LIFG, while more distant connections require greater control. Even though behavioural performance was maintained in response to TMS, LIFG stimulation increased the effect of semantic control demands in pMTG and pre-SMA, relative to stimulation of a control site (occipital pole). These changes were accompanied by reduced recruitment of both the stimulated region (LIFG) and its right hemisphere homologue (RIFG), particularly for strong associations with low control requirements. Thus repetitive TMS to LIFG modulated the contribution of distributed regions to semantic judgements in two distinct ways.

•

Offline rTMS was used to modulate the semantic control system.

•

fMRI revealed post-stimulation changes in other areas of the semantic control system.

•

Semantic retrieval requires the flexible activation of representations shaped by control processes.

Patterns of neural response in scene-selective regions of the human brain are affected by low-level manipulations of spatial frequency

Neuroimaging studies have found distinct patterns of response to different categories of scenes. However, the relative importance of low-level image properties in generating these response patterns is not fully understood. To address this issue, we directly manipulated the low level properties of scenes in a way that preserved the ability to perceive the category. We then measured the effect of these manipulations on category-selective patterns of fMRI response in the PPA, RSC and OPA. In Experiment 1, a horizontal-pass or vertical-pass orientation filter was applied to images of indoor and natural scenes. The image filter did not have a large effect on the patterns of response. For example, vertical- and horizontal-pass filtered indoor images generated similar patterns of response. Similarly, vertical- and horizontal-pass filtered natural scenes generated similar patterns of response. In Experiment 2, low-pass or high-pass spatial frequency filters were applied to the images. We found that image filter had a marked effect on the patterns of response in scene-selective regions. For example, low-pass indoor images generated similar patterns of response to low-pass natural images. The effect of filter varied across different scene-selective regions, suggesting differences in the way that scenes are represented in these regions. These results indicate that patterns of response in scene-selective regions are sensitive to the low-level properties of the image, particularly the spatial frequency content.

The neural correlates of semantic richness: evidence from an fMRI study of word learning

We investigated the neural correlates of concrete nouns with either many or few semantic features. A group of 21 participants underwent two days of training and were then asked to categorize 40 newly learned words and a set of matched familiar words as living or nonliving in an MRI scanner. Our results showed that the most reliable effects of semantic richness were located in the left angular gyrus (AG) and middle temporal gyrus (MTG), where activation was higher for semantically rich than poor words. Other areas showing the same pattern included bilateral precuneus and posterior cingulate gyrus. Our findings support the view that AG and anterior MTG, as part of the multimodal network, play a significant role in representing and integrating semantic features from different input modalities. We propose that activation in bilateral precuneus and posterior cingulate gyrus reflects interplay between AG and episodic memory systems during semantic retrieval.

Conceptual control across modalities: graded specialisation for pictures and words in inferior frontal and posterior temporal cortex

Controlled semantic retrieval to words elicits co-activation of inferior frontal (IFG) and left posterior temporal cortex (pMTG), but research has not yet established (i) the distinct contributions of these regions or (ii) whether the same processes are recruited for non-verbal stimuli. Words have relatively flexible meanings - as a consequence, identifying the context that links two specific words is relatively demanding. In contrast, pictures are richer stimuli and their precise meaning is better specified by their visible features - however, not all of these features will be relevant to uncovering a given association, tapping selection/inhibition processes. To explore potential differences across modalities, we took a commonly-used manipulation of controlled retrieval demands, namely the identification of weak vs. strong associations, and compared word and picture versions. There were 4 key findings: (1) Regions of interest (ROIs) in posterior IFG (BA44) showed graded effects of modality (e.g., words>pictures in left BA44; pictures>words in right BA44). (2) An equivalent response was observed in left mid-IFG (BA45) across modalities, consistent with the multimodal semantic control deficits that typically follow LIFG lesions. (3) The anterior IFG (BA47) ROI showed a stronger response to verbal than pictorial associations, potentially reflecting a role for this region in establishing a meaningful context that can be used to direct semantic retrieval. (4) The left pMTG ROI also responded to difficulty across modalities yet showed a stronger response overall to verbal stimuli, helping to reconcile two distinct literatures that have implicated this site in semantic control and lexical-semantic access respectively. We propose that left anterior IFG and pMTG work together to maintain a meaningful context that shapes ongoing semantic processing, and that this process is more strongly taxed by word than picture associations.

Neural mechanisms underlying song and speech perception can be differentiated using an illusory percept

The issue of whether human perception of speech and song recruits integrated or dissociated neural systems is contentious. This issue is difficult to address directly since these stimulus classes differ in their physical attributes. We therefore used a compelling illusion (Deutsch et al. 2011) in which acoustically identical auditory stimuli are perceived as either speech or song. Deutsch's illusion was used in a functional MRI experiment to provide a direct, within-subject investigation of the brain regions involved in the perceptual transformation from speech into song, independent of the physical characteristics of the presented stimuli. An overall differential effect resulting from the perception of song compared with that of speech was revealed in right midposterior superior temporal sulcus/right middle temporal gyrus. A left frontotemporal network, previously implicated in higher-level cognitive analyses of music and speech, was found to co-vary with a behavioural measure of the subjective vividness of the illusion, and this effect was driven by the illusory transformation. These findings provide evidence that illusory song perception is instantiated by a network of brain regions that are predominantly shared with the speech perception network.

A robust implementation of a kurtosis beamformer for the accurate identification of epileptogenic foci

OBJECTIVE

To improve the accuracy and reliability of the localisation of epileptogenic activity using spatially filtered MEG data.

METHODS

A synthetic epileptic source was embedded in healthy brain activity in different orientations in order to estimate how reliably this signal containing high levels of kurtosis can be localised. An existing approach (SAM(g2)) was compared to a new implementation of the methodology.

RESULTS

The results confirm that a kurtosis beamformer is an effective tool with which to localise spontaneous epileptiform activity. However, it is crucial that the orientation of source reconstruction matches that of the true source otherwise the epileptic activity is either mis-localised or completely missed. Therefore as the original SAM(g2) implementation is restricted to the tangential plane, in certain circumstances it will perform poorly compared to the approach described here.

CONCLUSIONS

A kurtosis beamformer is made more accurate and more robust if the analysis is not restricted to the tangential plane and if the optimisation routine for selecting the source orientation is performed using kurtosis rather than power.

SIGNIFICANCE

MEG is increasingly being used for the non-invasive localisation of epileptic biomagnetic signals and the implementation described in this paper increases the clinical utility of the technique.

Examining the effects of one- and three-dimensional spatial filtering analyses in magnetoencephalography

Spatial filtering, or beamforming, is a commonly used data-driven analysis technique in the field of Magnetoencephalography (MEG). Although routinely referred to as a single technique, beamforming in fact encompasses several different methods, both with regard to defining the spatial filters used to reconstruct source-space time series and in terms of the analysis of these time series. This paper evaluates two alternative methods of spatial filter construction and application. It demonstrates how encoding different requirements into the design of these filters has an effect on the results obtained. The analyses presented demonstrate the potential value of implementations which examine the timeseries projections in multiple orientations at a single location by showing that beamforming can reconstruct predominantly radial sources in the case of a multiple-spheres forward model. The accuracy of source reconstruction appears to be more related to depth than source orientation. Furthermore, it is shown that using three 1-dimensional spatial filters can result in inaccurate source-space time series reconstruction. The paper concludes with brief recommendations regarding reporting beamforming methodologies in order to help remove ambiguity about the specifics of the techniques which have been used.

Negative BOLD in the Lateral Geniculate Nucleus: Neuronal Implications and Cortico-Thalamic Feedback

Previous research has demonstrated a sustained negative BOLD response (NBR) that is negatively correlated with the spatio-temporal properties of a visual stimulus. Whilst it has been suggested that the NBR surrounding the positive BOLD response (PBR) may reflect blood-stealing, evidence indicates that the extensive NBR distal to the PBR is a manifestation of neuronal suppression. This study aimed to evaluate NBR in the lateral geniculate nucleus (LGN), and to explore the source of the NBR. fMRI data were obtained from six subjects, while they viewed a grating stimulus. The NBR was identified in the LGN ipsilateral to the stimulus. The results also verified the NBR in V1 ipsilateral to the stimulus and revealed the PBR in bilateral V5. It was concluded that the NBR can be found in the LGN, and is most likely driven by feedback from ipsilateral V1. The finding that the stimulus that stimulates the LGN in one hemisphere can cause extensive suppression in the LGN of the opposite hemisphere rejects the notion that the effect is purely a blood-stealing effect as the two LGN have different blood supplies. The results, together with previous research, indicate that the NBR may reflect neuronal suppression.

Non-parametric statistical thresholding of baseline free MEG beamformer images

Magnetoencephalography (MEG) provides excellent temporal resolution when examining cortical activity in humans. Inverse methods such as beamforming (a spatial filtering approach) provide the means by which activity at cortical locations can be estimated. To date, the majority of work in this field has been based upon power changes between active and baseline conditions. Recent work, however, has focused upon other properties of the time series data reconstructed by these methods. One such metric, the Source Stability Index (SSI), relates to the consistency of the time series calculated only over an active period without the use of a baseline condition. In this paper we apply non-parametric statistics to SSI volumetric maps of simulation, auditory and somatosensory data in order to provide a robust and principled method of statistical inference in the absence of a baseline condition.