Publications

{126 human brain field maps from 18 normal subjects in 7 head positions each are published. The data may be of interest for shim coil designers as well as fro susceptibility mapping research.}

{Multi-coil shim setup is a popular choice for B0 shimming. In contrast to conventional regular arrangement of the shim coils, one can effectively position the shim coil to match inhomogeneity distribution in the human brain. In this work, a comparison between regular and optimized arrangement of the local coils in a multi-coil shim setup is performed and the pros and cons of each design are evaluated.}

{Purpose: We designed and implemented an orthogonal shim array consisting of shim coils with their planes perpendicular to the planes of neighboring RF coils. This shim coil improves the magnetic field homogeneity by minimizing the interference to RF coils. Methods Using realistic off-resonance maps of the human brain, we first evaluated the performance of shim coils in different orientations. Based on simulations, we developed a 7-channel orthogonal shim array, whose coil plan was perpendicular to neighboring RF coils, at the forehead. A programmable open-source current driver supplied shim currents. Results The 7-channel orthogonal shim array caused only marginal SNR loss to the integrated 32-channel RF-shim array. The 7-channel orthogonal shim array itself improved the magnetic field homogeneity by 30\textpercent in slice-optimized shimming, comparable to the baseline shimming offered by the scanner\textquotesingles 2nd order spherical harmonic shimming. Conclusion Orthogonal shim coils can improve the field homogeneity while maintaining high image SNR.}

{The following slides presents an analysis of the spread of COVID-19 inGermany, and the effects of different containment strategies.Details of the mathematical model and how the parameters of themodel are estimated from the available data on the spread ofCOVID-19 is explained in:https://people.tuebingen.mpg.de/vbokharaie/Estimating\textunderscoreCovid19\textunderscorecontact\textunderscorerates.pdfThe following few slide provides enough information to make thepresented figures understandable.}

{Laughter is a multifaceted signal, which can convey social acceptance facilitating social bonding as well as social rejection inflicting social pain. In the current study, we addressed the neural correlates of social intent attribution to auditory or visual laughter within an fMRI study to identify brain areas showing linear increases of activation with social intent ratings. Negative social intent attributions were associated with activation increases within the medial prefrontal cortex/anterior cingulate cortex (mPFC/ACC). Interestingly, negative social intent attributions of auditory laughter were represented more rostral than visual laughter within this area. Our findings corroborate the role of the mPFC/ACC as key node for processing \textquotedblleftsocial pain\textquotedblright with distinct modality-specific subregions. Other brain areas that showed an increase of activation included bilateral inferior frontal gyrus and right superior/middle temporal gyrus (STG/MTG) for visually presented laughter and bilateral STG for auditory presented laughter with no overlap across modalities. Similarly, positive social intent attributions were linked to hemodynamic responses within the right inferior parietal lobe and right middle frontal gyrus, but there was no overlap of activity for visual and auditory laughter. Our findings demonstrate that social intent attribution to auditory and visual laughter is located in neighboring, but spatially distinct neural structures.}

{An international group of researchers met in November 2019 in Beijing to explore the intersection of neuroscience and AI. The aim was to offer a fertile ground for stimulating discussions and ideas, including issues such as policy making and the future of neuroscience and AI across the globe.}

{Large-scale testing is considered key to assessing the state of the current COVID-19 pandemic, yet interpreting such data remains elusive. We modeled competing hypotheses regarding the underlying testing mechanisms, thereby providing different prevalence estimates based on case numbers, and used them to predict SARS-CoV-2-attributed death rate trajectories. Assuming that individuals were tested based solely on a predefined risk of being infectious implied the absolute case numbers reflected prevalence, but turned out to be a poor predictor. In contrast, models accounting for testing capacity, limiting the pool of tested individuals, performed better. This puts forward the percentage of positive tests as a robust indicator of epidemic dynamics in absence of country-specific information. We next demonstrated this strongly affects data interpretation. Notably absolute case numbers trajectories consistently overestimated growth rates at the beginning of two COVID-19 epidemic waves. Overall, this supports non-trivial testing mechanisms can be inferred from data and should be scrutinized.}

{We developed a novel 9.4T (400MHz) human head transceiver array consisted of 8 optimized bent folded dipole antennas. Due to an asymmetrical shape of dipoles (bending) and the RF shield, the array simultaneously excites two modes including a circular polarized mode of the array itself, and the TE mode of the human head. Mode mixing can be easily controlled by changing the folded length. As a result, the new array provides superior whole-brain coverage compared to various 8-element loop and dipole arrays or even to a more complicated 16-element loop array. In addition, the maximum local SAR is substantially reduced.}

{Animated virtual characters are essential to many applications. Little is known so far about biological and personality inferences made from a virtual character\textquoterights body shape and motion. Here, we investigated how sex-specific differences in walking style relate to the perceived attractiveness and confidence of male and female virtual characters. The characters were generated by reconstructing body shape and walking motion from optical motion capture data. The results suggest that sexual dimorphism in walking style plays a different role in attributing biological and personality traits to male and female virtual characters. This finding has important implications for virtual character animation.}

{The hippocampus has a major role in encoding and consolidating long-term memories, and undergoes plastic changes during sleep1. These changes require precise homeostatic control by subcortical neuromodulatory structures2. The underlying mechanisms of this phenomenon, however, remain unknown. Here, using multi-structure recordings in macaque monkeys, we show that the brainstem transiently modulates hippocampal network events through phasic pontine waves known as pontogeniculooccipital waves (PGO waves). Two physiologically distinct types of PGO wave appear to occur sequentially, selectively influencing high-frequency ripples and low-frequency theta events, respectively. The two types of PGO wave are associated with opposite hippocampal spike-field coupling, prompting periods of high neural synchrony of neural populations during periods of ripple and theta instances. The coupling between PGO waves and ripples, classically associated with distinct sleep stages, supports the notion that a global coordination mechanism of hippocampal sleep dynamics by cholinergic pontine transients may promote systems and synaptic memory consolidation as well as synaptic homeostasis.}

{Introduction: Previously we proposed a supervised learning approach to automatically generate MR sequences from scratch without providing sequence programming rules, called MRzero [1]. In the present work, we develop an auto-encoder for T1 by performing a joint optimization of sequence parameters and a neural network using MRzero. Subjects/methods: The fully differentiable MRI pipeline is simulated end-to-end with Bloch parameters as input and T1 as target. We utilize known operator learning [2] in the reconstruction to reduce the number of trainable parameters in our NN by keeping the adjoint formalism [1] as known operator in the image reconstruction. The T1 training dataset consist of ten T1 maps with matrix size 32 9 32. For each target sample a non-zero PD rectangle with matrix size 16 9 16 at varying spatial location with voxel-wise randomly assigned PD, T1, T2 and B0 is defined, resulting in a total training data size of 2560 samples. A three-hidden-layer multilayer perceptron is used for T1 quantification. The MR sequence is based on a 180 deg inversion prepared 2D FLASH sequence with matrix size 32 9 32, TR \textequals 15 ms, TE \textequals 8 ms, FA \textequals 5 deg, repeated 6 times with varying TI and Trec. Together with the NN parameters, all TI and Trec times are optimized to find the best sequence for T1 mapping and are initialized with 0. Additionally, a penalty for longer times was applied to enforce shorter sequences. The optimization process (Fig. 1) interleaves the sequence and NN optimization after 50 and 5000 iterations, respectively. In total 500 iterations of sequence optimization are performed. Simultaneously optimized sequence parameters and trained NN are applied on a higher resolution with matrix size 126 9 126 and parallel imaging (GRAPPA acceleration factor 3) for in vivo measurements at 3T. Results/discussion: The T1 map of a healthy subject generated by the final optimized sequence is displayed in Fig. 2. Figure 3 shows the different stages of sequence optimization. The acquired T1 values of CSF, white matter and grey matter for later iterations match well to literature values at 3T [3]. A standard inversion recovery sequence was used as reference. The obtained maps match well with the reference, but the acquisition time could be reduced from 63.3 s to 19.2 s. Optimized TI and Trec times range from 0.5 s to 1.8 s and 0.5 s to 1.1 s, respectively. The simultaneous sequence optimization and NN training was performed solely on synthetic data at low resolution, but inference on higher resolution on in vivo data provided high quality T1 maps. Preliminary results at low resolution were shown in [1]. The T1 autoencoder is a proof-of-concept that can be extended also to multiparametric mapping\textemdashsimilar to MR fingerprinting\textemdashyielding PD, T1, and T2, as well as B1 and B0 inhomogeneity maps.}

{We use a UTE sequence combining with Dixon method to obtain the subject specific susceptibility distribution, with 4-class tissue segmentation. The susceptibility model was then used to simulate motion-induced B0 change for two head positions. A good agreement between the simulated and measured field map has been observe. A forward field map predicting strategy was explored using the susceptibility model.}

{Brain derived neurotrophic factor (BDNF) has been implicated in the pathophysiology of schizophrenia. As BDNF regulates axonal and dendritic growth, altered BDNF levels in schizophrenia patients might underlie changes in structural connectivity that have been identified by magnetic resonance imaging (MRI). We investigated a possible correlation between BDNF serum levels, fiber tract architecture, and regional grey matter volumes in 19 schizophrenia patients and a gender- and age-matched control group. Two patients had to be excluded due to abnormalities in their MRI scans. Serum samples were obtained to determine BDNF levels, and T1- as well as diffusion-weighted sequences were acquired. We, then, investigated correlations between BDNF serum levels with neuroimaging parameters, using Voxel-based Morphometry (VBM) and Tract-based Spatial Statistics (TBSS). We found a significant negative correlation between BDNF serum levels and FA values in the right inferior fronto-occipital fasciculus and the right superior longitudinal fasciculus. These regions also showed a decrease in AD values in schizophrenia patients. Grey matter volumes were reduced in patients but there was no correlation between regional grey matter volumes and BDNF. The right superior longitudinal fasciculus has been repeatedly identified to exhibit microstructural changes in schizophrenia patients. Our findings of a negative correlation between BDNF and FA values in patients might indicate that BDNF is upregulated to compensate decreased structural connectivity as it induces neural plasticity and shows increased levels in damaged tissue. These findings of our pilot study are encouraging leads for future research in larger samples.}

{We obtain an explicit formula for the full expansion of the spectral action on Robertson\textendashWalker spacetimes, expressed in terms of Bell polynomials, using Brownian bridge integrals and the Feynman\textendashKac formula. We then apply this result to the case of multifractal Packed Swiss Cheese Cosmology models obtained from an arrangement of Robertson\textendashWalker spacetimes along an Apollonian sphere packing. Using Mellin transforms, we show that the asymptotic expansion of the spectral action contains the same terms as in the case of a single Robertson\textendashWalker spacetime, but with zeta-regularized coefficients, given by values at integers of the zeta function of the fractal string of the radii of the sphere packing, as well as additional log-periodic correction terms arising from the poles (off the real line) of this zeta function.}

{Purpose: To provide transmit whole-brain coverage at 9.4 T using an array with only eight elements and improve the specific absorption rate (SAR) performance, a novel dipole array was developed, constructed, and tested. Methods: The array consists of eight optimized bent folded-end dipole antennas circumscribing a head. Due to the asymmetrical shape of the dipoles (bending and folding) and the presence of an RF shield near the folded portion, the array simultaneously excites two modes: a circular polarized mode of the array itself, and the TE mode ("dielectric resonance") of the human head. Mode mixing can be controlled by changing the length of the folded portion. Due to this mixing, the new dipole array improves longitudinal coverage as compared with unfolded dipoles. By optimizing the length of the folded portion, we can also minimize the peak local SAR (pSAR) value and decouple adjacent dipole elements. Results: The new array improves the SEE (\textless B+1 \textgreater/$\surd$pSAR) value by about 50\textpercent, as compared with the unfolded bent dipole array. It also provides better whole-brain coverage compared with common single-row eight-element dipole arrays, or even to a more complex double-row 16-element surface loop array. Conclusion: In general, we demonstrate that rather than compensating for the constructive interference effect using additional hardware, we can use the "dielectric resonance" to improve coverage, transmit field homogeneity, and SAR efficiency. Overall, this design approach not only improves the transmit performance in terms of the coverage and SAR, but substantially simplifies the common surface loop array design, making it more robust, and therefore safer.}

{Introduction: Here we present a new approach called \textasciigrave\textasciigraveBipartite connectivity mapping (BCM)\textquotesingle\textquotesingle for analyzing connectivity in rs-fMRI between two brain regions. The two regions can be of any size, but should be non-overlapping. The main idea is to represent connectivity between the regions as a bipartite graph and analyse it using bipartite network projections. The advantage of this approach is that it allows to fully preserve spatial precision in both ROIs. This sets it apart from traditional seed-based connectivity mapping where the seed ROI must be averaged so that its spatial information is lost. Here we show an application of this idea to rs-fMRI data acquired at 9.4 Tesla. Methods: A bipartite graph is a graph whose edges have endpoints connecting the two regions, but that do not connect vertices within the same region. Our proposed method BCM is a general framework that offers several options for analysing bipartite graphs. The first and easiest technique is to compute the hubs of the bipartite graph using either degree centrality or eigenvector centrality mapping [1] and map them to each voxel. The second technique is to project the bipartite graph onto one of the ROIs and compute the hubs of the projected network. There are several choices for projections of bipartite graphs [2], one of which is currently implemented in BCM. Results: Resting state fMRI data of a healthy female subject were acquired at a 9.4 Tesla scanner (Siemens). A gradient echo EPI sequence with TR\textequals2.03 sec, (1.2)\textasciicircum3 mm isotropic resolution was used. The scan time was 12 minutes (rs-fMRI with eyes open). Preprocessing consisted of corrections for motion and removal of base line drifts via a highpass filter. Two ROIs were manually defined with one ROI (A) corresponding to left and right thalamus, and ROI B corresponding to right-hemispheric grey matter. The size of A was 10618 voxels, the size of B was 120209 voxels. The hubs of the bipartite graph are shown below (Figure 1). The BCM results in the two ROIs are scaled independently for better visualization. Note that LGN shows up prominently in A, while V1 is clearly visible in B. Using the same data as above, we applied a BCM network projection. Here we projected the above bipartite graph onto ROI A (thalamus). Below (Figure 2) we show the degree centrality mapping of the projected network. Conclusions: The new BCM algorithm provides a method for analyzing connectivity between two ROIs. It can for instance be used for detecting nuclei in subcortical regions that are characterized by their connectivity to other brain areas. BCM is a general framework for handling bipartite graphs. At present, two definitions of hubs are implemented, namely degree centrality and eigenvector centrality. Furthermore, a method for network projection is implemented. We believe that BCM offers an alternative route for ROI-based connectivity analysis in fMRI. Its main advantage is that it preserves spatial specificity in both ROIs.}

{Correct calibration of the transmit field B1+ is crucial to achieve optimal SNR. However, fast and robust B1+ calibration is difficult for X-nuclei due to the low signal sensitivity. In this work, we proposed a fast B1+ calibration method based on the Bloch-Siegert shift and single voxel ISIS localization. With the proposed sequence, the B1+ calibration can be done in less than 5 min for the human brain at B0 \textequals 9.4 T.}

{Retinal photoreceptors modulate the pupil diameter to regulate retinal illumination. At early stage the pupil-response is formed by intrinsically-photosensitive-Retinal-Ganglion-Cells (ipRGCs) expressing melanopsin, activated by blue light. ipRGCs\textquoteright axons pass through the optic nerve head, corresponding to the blind-spot. No photoreceptors except melanopsin appear to exist in the blind-spot. Contributions of melanopsin to pupil constriction in absence of classical photoreceptors in the blind-spot is not fully understood. We investigated how blue light in the blind-spot changes melanopsin-pupil-response compared to parafovea and periphery. The Post-Illumination-Pupil-Response (PIPR) amplitude reflecting melanopsin was analyzed for standardized time windows (1s\textless1.7s, 1s\textgreater1.8s and 2\textendash6s) and expressed as pupillary-change. Bayesian analysis showed a BF\textgreater3 that PIPR\textgreater1.8s for blind-spot and periphery is not different. At times 2s\textendash6s, a t-test comparison in the blind-spot condition showed a sign ificantly larger PIPR to blue compared to red light, confirming a melanopsin-pupil-response in the blind-spot. Taken together, equivalent stimulation in the blind-spot and periphery revealed comparable PIPR, although there are no rods and cones in the blind-spot. In absence of classical photoreceptors in the blind-spot, melanopsin seems to be responsible for pupil constriction in similar manner as in the periphery, which supports the presence of melanopsin on the axons of ipRGCs.}

{It has been difficult to find robust brain structural correlates of the overall severity of major depressive disorder (MDD). We hypothesized that specific symptoms may better reveal correlates and investigated this for the severity of insomnia, both a key symptom and a modifiable major risk factor of MDD. Cortical thickness, surface area and subcortical volumes were assessed from T1-weighted brain magnetic resonance imaging (MRI) scans of 1053 MDD patients (age range 13-79 years) from 15 cohorts within the ENIGMA MDD Working Group. Insomnia severity was measured by summing the insomnia items of the Hamilton Depression Rating Scale (HDRS). Symptom specificity was evaluated with correlates of overall depression severity. Disease specificity was evaluated in two independent samples comprising 2108 healthy controls, and in 260 clinical controls with bipolar disorder. Results showed that MDD patients with more severe insomnia had a smaller cortical surface area, mostly driven by the right insula, left inferior frontal gyrus pars triangularis, left frontal pole, right superior parietal cortex, right medial orbitofrontal cortex, and right supramarginal gyrus. Associations were specific for insomnia severity, and were not found for overall depression severity. Associations were also specific to MDD; healthy controls and clinical controls showed differential insomnia severity association profiles. The findings indicate that MDD patients with more severe insomnia show smaller surfaces in several frontoparietal cortical areas. While explained variance remains small, symptom-specific associations could bring us closer to clues on underlying biological phenomena of MDD.}

{Purpose To understand the influence of various acquisition parameters on the ability of CEST MR-Fingerprinting (MRF) to discriminate different chemical exchange parameters and to provide tools for optimal acquisition schedule design and parameter map reconstruction. Methods Numerical simulations were conducted using a parallel computing implementation of the Bloch-McConnell equations, examining the effect of TR, TE, flip-angle, water urn:x-wiley:07403194:media:mrm27937:mrm27937-math-0011 and urn:x-wiley:07403194:media:mrm27937:mrm27937-math-0012, saturation-pulse duration, power, and frequency on the discrimination ability of CEST-MRF. A modified Euclidean distance matching metric was evaluated and compared to traditional dot product matching. L-Arginine phantoms of various concentrations and pH were scanned at 4.7T and the results compared to numerical findings. Results Simulations for dot product matching demonstrated that the optimal flip-angle and saturation times are urn:x-wiley:07403194:media:mrm27937:mrm27937-math-0013 and 1100 ms, respectively. The optimal maximal saturation power was 3.4 $\mu$T for concentrated solutes with a slow exchange rate, and 5.2 $\mu$T for dilute solutes with medium-to-fast exchange rates. Using the Euclidean distance matching metric, much lower maximum saturation powers were required (1.6 and 2.4 $\mu$T, respectively), with a slightly longer saturation time (1500 ms) and urn:x-wiley:07403194:media:mrm27937:mrm27937-math-0014 flip-angle. For both matching metrics, the discrimination ability increased with the repetition time. The experimental results were in agreement with simulations, demonstrating that more than a 50\textpercent reduction in scan-time can be achieved by Euclidean distance-based matching. Conclusions Optimization of the CEST-MRF acquisition schedule is critical for obtaining the best exchange parameter accuracy. The use of Euclidean distance-based matching of signal trajectories simultaneously improved the discrimination ability and reduced the scan time and maximal saturation power required.}

{The inference of causal relationships using observational data from partially observed multivariate systems with hidden variables is a fundamental question in many scientific domains. Methods extracting causal information from conditional independencies between variables of a system are common tools for this purpose, but are limited in the lack of independencies. To surmount this limitation, we capitalize on the fact that the laws governing the generative mechanisms of a system often result in substructures embodied in the generative functional equation of a variable, which act as sufficient statistics for the influence that other variables have on it. These functional sufficient statistics constitute intermediate hidden variables providing new conditional independencies to be tested. We propose to use the Information Bottleneck method, a technique commonly applied for dimensionality reduction, to find underlying sufficient sets of statistics. Using these statistics we formulate new additional rules of causal orientation that provide causal information not obtainable from standard structure learning algorithms, which exploit only conditional independencies between observable variables. We validate the use of sufficient statistics for structure learning both with simulated systems built to contain specific sufficient statistics and with benchmark data from regulatory rules previously and independently proposed to model biological signal transduction networks.}

{Goal-directed behavior requires the interaction of multiple brain regions. How these regions and their interactions with brain-wide activity drive action selection is less understood. We have investigated this question by combining whole-brain volumetric calcium imaging using light-field microscopy and an operant-conditioning task in larval zebrafish. We find global, recurring dynamics of brain states to exhibit pre-motor bifurcations toward mutually exclusive decision outcomes. These dynamics arise from a distributed network displaying trial-by-trial functional connectivity changes, especially between cerebellum and habenula, which correlate with decision outcome. Within this network the cerebellum shows particularly strong and predictive pre-motor activity (\textgreater10 s before movement initiation), mainly within the granule cells. Turn directions are determined by the difference neuroactivity between the ipsilateral and contralateral hemispheres, while the rate of bi-hemispheric population ramping quantitatively predicts decision time on the trial-by-trial level. Our results highlight a cognitive role of the cerebellum and its importance in motor planning.}

{To determine own upright body orientation the brain creates a sense of verticality by a combination of multisensory inputs. To test whether this process is affected by aging, we placed younger and older adults on a motion platform and systematically tilted the orientation of their visual surroundings by using an augmented reality setup. In a series of trials, participants adjusted the orientation of the platform until they perceived themselves to be upright. Tilting the visual scene around the roll axis induced a bias in subjective postural vertical determination in the direction of scene tilt in both groups. In the group of older participants, however, the observed peak bias was larger and occurred at larger visual tilt angles. This indicates that the susceptibility to visually induced biases increases with age, possibly caused by a reduced reliability of sensory information.}

{We developed a nanosized perfluorocarbon-based system with incorporated paramagnetic Gd(iii) chelates, able to generate a quantitative 19F MRI signal, while in parallel capable of modulating the 1H MRI signal in response to the coordination of Ca2+ ions. Subsequently, we performed experiments in vivo and estimated the concentration of the applied probe in the tissue by means of 19F MRI.}

{How do people perceive and communicate structure? We investigate this question by letting participants play a communication game, where one player describes a pattern, and another player redraws it based on the description alone. We use this paradigm to compare two models of pattern description, one compositional (complex structures built out of simpler ones) and one non-compositional. We find that compositional patterns are communicated more effectively than non-compositional patterns, that a compositional model of pattern description predicts which patterns are harder to describe, and that this model can be used to evaluate participants\textquoteright drawings, producing human-like quality ratings. Our results suggest that natural language can tap into a compositionally structured pattern description language.}

{Computational psychiatry is a nascent field that attempts to use multi-level analyses of the underlying computational problems that we face in navigating a complex, uncertain and changing world to illuminate mental dysfunction and disease. Two particular foci of the field are the costs and benefits of environmental adaptivity and the danger and necessity of heuristics. Here, we examine the extent to which these foci and others can be used to study the actual and potential flaws of the artificial computational devices that we are increasingly inventing and empowering to navigate this very same environment on our behalf.}

{The insula has been classically divided into vast granular, dysgranular and agranular sectors. Over the years, several distinct studies proposed subdivisions of these sectors, with however no consensus. We recently proposed a cyto- and myelo-architectonic partition in which each sector contained sharply delimited areas (Evrard et al. 2014 J Comp Neurol 522: 64-97). Some of these areas were further divided into distinct subareas with obvious functional implications. Here, we examined the spatial relationship between architectonic boundaries and tract-tracing labeling in the insula in the macaque monkey. Injections of neuronal tracers in distinct areas of the prefrontal or anterior cingulate cortices produced heterogeneous and discontinuous patterns of anterograde and retrograde labeling in the insula. These patterns were made of sharply delimited patches forming anteroposterior stripes across consecutive coronal sections. While the overall pattern of labeling varied with the injection site, the patches systematically coincided with specific architectonic subareas, particularly in the dysgranular insula. This unequivocally validates our prior architectonic partition and strongly supports the idea of a refined modular Bauplan of the primate insula. This modular organization may underlie a serial stream of integration of interoception with \textquoteleftself-agency\textquoteright and \textquoteleftsocial\textquoteright activities across distinct insulo-prefrontal processing units that need to be explored.}

{Radio-frequency (rf) coils are used in all clinical and research magnetic-resonance-imaging (MRI) systems to excite nuclear spins and to receive signals from them. The quality of imaging depends strongly on the signal-to-noise ratio (SNR) and the transmit efficiency of the coils. The birdcage volume coils used in most MRI scanners for homogeneous imaging of a sample are typically shielded from the external systems of the MRI scanner, i.e., the gradient coils, to confine the rf field within the region of interest. However, the near magnetic field of a conventional copper rf shield surrounding a birdcage coil interferes destructively with the primary field of the coil in the sample, which significantly limits the SNR and the transmit efficiency. In the work presented here, we theoretically study and experimentally demonstrate the possibility of creating an artificial magnetic rf shield for a birdcage coil with constructive interference in a sample. This effect is similar to the in-phase reflection of antenna far fields from a magnetic shield but affects the near field in MRI. We build an analytical model of a birdcage coil shielded with a cylindrical impedance boundary and analyze the conditions for increasing the efficiency of the coil by means of the shield. We conclude that by replacing a copper shield with an artificial magnetic one, it is possible to reduce the dissipative intrinsic losses of the coil and increase the power absorbed by the sample, which improves the efficiency. To demonstrate the effect, we perform a detailed numerical simulation and an experiment with a small birdcage in a 7-T 19F MRI system with a magnetic shield implemented as a periodic cylindrical metal structure with corrugations filled with a ceramic.}

{Even when we are wearing gloves, we can easily detect whether a surface that we are touching is sticky or not. However, we know little about the similarities between brain activations elicited by this glove contact and by direct contact with our bare skin. In this functional magnetic resonance imaging (fMRI) study, we investigated which brain regions represent stickiness intensity information obtained in both touch conditions, i.e., skin contact and glove contact. First, we searched for neural representations mediating stickiness for each touch condition separately and found regions responding to both mainly in the supramarginal gyrus and the secondary somatosensory cortex. Second, we explored whether surface stickiness is encoded in common neural patterns irrespective of how participants touched the sticky stimuli. Using a cross-condition decoding method, we tested whether the stickiness intensities could be decoded from fMRI signals evoked by skin contact using a classifier trained on the responses elicited by glove contact, and vice versa. Our results found shared neural encoding patterns in the bilateral angular gyri and the inferior frontal gyrus (IFG) and suggest that these areas represent stickiness intensity information regardless of how participants touched the sticky stimuli. Interestingly, we observed that neural encoding patterns of these areas were reflected in participants\textquoteright intensity ratings. This study revealed common and distinct brain activation patterns of tactile stickiness using two different touch conditions, which may broaden the understanding of neural mechanisms related to surface texture perception.}

{Deep generative models can emulate the perceptual properties of complex image datasets, providing a latent representation of the data. However, manipulating such representation to perform meaningful and controllable transformations in the data space remains challenging without some form of supervision. While previous work has focused on exploiting statistical independence to \textit\textbraceleftdisentangle\textbraceright latent factors, we argue that such requirement can be advantageously relaxed and propose instead a non-statistical framework that relies on identifying a modular organization of the network, based on counterfactual manipulations. Our experiments support that modularity between groups of channels is achieved to a certain degree on a variety of generative models. This allowed the design of targeted interventions on complex image datasets, opening the way to applications such as computationally efficient style transfer and the automated assessment of robustness to contextual changes in pattern recognition systems.}

{Multiple theories attribute to the primate prefrontal cortex a critical role in conscious perception. However, opposing views caution that prefrontal activity could reflect other cognitive variables during paradigms investigating consciousness, such as decision-making, monitoring and motor reports. To resolve this ongoing debate, we recorded from prefrontal ensembles of macaque monkeys during a no-report paradigm of binocular rivalry that instigates internally driven transitions in conscious perception. We could decode the contents of consciousness from prefrontal ensemble activity during binocular rivalry with an accuracy similar to when these stimuli were presented without competition. Oculomotor signals, used to infer conscious content, were not the only source of these representations since visual input could be significantly decoded when eye movements were suppressed. Our results suggest that the collective dynamics of prefrontal cortex populations reflect internally generated changes in the content of consciousness during multistable perception.}

{Dipole antennas have been successfully utilized at ultra-high fields (UHF, \textgreater7 T) as elements of human body arrays. Usage of dipoles for UHF human head arrays is still under development. In this case, dipoles must be made much shorter, and placed at a relatively large distance to the head. As a result, dipoles are not well loaded and are often purely decoupled. In this work, we developed a novel method of decoupling of adjacent dipole antennas, and used this technique while constructing a novel 9.4 T human head TxRx dipole array coil. The array demonstrates good decoupling and full-brain coverage.}

{Dipole antennas have recently been introduced to the field of MRI and successfully used, mostly as elements of ultra-high field (UHF, $\geq$ 7 T) human body arrays. Usage of dipole antennas for UHF human head transmit (Tx) arrays is still under development. Due to the substantially smaller size of the sample, dipoles must be made significantly shorter than in the body array. Additionally, head Tx arrays are commonly placed on the surface of rigid helmets made sufficiently large to accommodate tight-fit receive arrays. As a result, dipoles are not well loaded and are often poorly decoupled, which compromises Tx efficiency. Commonly, adjacent array elements are decoupled by circuits electrically connected to them. Placement of such circuits between distantly located dipoles is difficult. Alternatively, decoupling is provided by placing passive antennas between adjacent dipole elements. This method only works when these additional components are sufficiently small (compared with the size of active dipoles). Otherwise, RF fields produced by passive elements interfere destructively with the RF field of the array itself, and previously reported designs have used passive dipoles of about the size of array dipoles. In this work, we developed a novel method of decoupling for adjacent dipole antennas, and used this technique while constructing a 9.4 T human head eight-element transceiver array. Decoupling is provided without any additional circuits by simply folding the dipoles and using an RF shield located close to the folded portion of the dipoles. The array reported in this work demonstrates good decoupling and whole-brain coverage.}

{Oncolytic virotherapy (OV) is a promising treatment for high mortality cancers.1Non-invasive monitoring of OV is essential for optimizing the clinical outcome and providing an improved understanding of the interactions between the virus and its tumor-host. Chemical exchange saturation transfer (CEST) MRI is a molecular imaging technique that may shed new light on OV, as it is capable of detecting pro-tein concentration and pH changes. However, clinical translation of CEST methods has been hindered by the qualitative nature of the image contrast and the long image acquisition times.}

{Prominent asymmetries in the bSSFP frequency profile in tissues with distinct fiber pathways are known to be a confounding factor in the quantification of relaxation times from a series of phase-cycled scans. It has been demonstrated that the resulting bias can be eliminated by training artificial neural networks using gold standard relaxation times as target. Here, the ability of neural networks to not only provide gold standard brain tissue T1 and T2 values as well as field map estimates (B1, $\Delta$B0) but also to highly accelerate the acquisition by reducing the number of phase-cycles is explored.}

{Purpose: Calculation of sophisticated MR contrasts often requires complex mathematical modeling. Data evaluation is computationally expensive, vulnerable to artifacts, and often sensitive to fit algorithm parameters. In this work, we investigate whether neural networks can provide not only fast model fitting results, but also a quality metric for the predicted values, so called uncertainty quantification, investigated here in the context of multi-pool Lorentzian fitting of CEST MRI spectra at 3T. Methods A deep feed-forward neural network including a probabilistic output layer allowing for uncertainty quantification was set up to take uncorrected CEST-spectra as input and predict 3T Lorentzian parameters of a 4-pool model (water, semisolid MT, amide CEST, NOE CEST), including the B0 inhomogeneity. Networks were trained on data from 3 subjects with and without data augmentation, and applied to untrained data from 1 additional subject and 1 brain tumor patient. Comparison to conventional Lorentzian fitting was performed on different perturbations of input data. Results The deepCEST 3T networks provided fast and accurate predictions of all Lorentzian parameters and were robust to input perturbations because of noise or B0 artifacts. The uncertainty quantification detected fluctuations in input data by increase of the uncertainty intervals. The method generalized to unseen brain tumor patient CEST data. Conclusions The deepCEST 3T neural network provides fast and robust estimation of CEST parameters, enabling online reconstruction of sophisticated CEST contrast images without the typical computational cost. Moreover, the uncertainty quantification indicates if the predictions are trustworthy, enabling confident interpretation of contrast changes.}

{Analysis of CEST data often requires complex mathematical modeling before contrast generation, which can be error prone and time-consuming. Here, a probabilistic deep learning approach is introduced to shortcut conventional Lorentzian fitting analysis of 3T in-vivo CEST data by learning from previously evaluated data. It is demonstrated that the trained networks generalize to data of a healthy subject and a brain tumor patient, providing CEST contrasts in a fraction of the conventional evaluation time. Additionally, the probabilistic network architecture enables uncertainty quantification, indicating if predictions are trustworthy, which is assessed by perturbation analysis.}

{Different classes of small- or nano-sized calcium-sensitive probes for magnetic resonance imaging (MRI) have been proposed in the last two decades. These compounds have been developed mainly for functional MRI purposes and tested in vivo in different animal models. Most of them are paramagnetic systems that change their relaxivity in the presence of the divalent ion calcium, resulting in increased T1 or T2 contrast. In this work, we report the investigation of their relaxometric behavior at low magnetic fields, specifically the comparison of the monomeric Ca-sensitive probe and the corresponding dendrimer conjugates of generations 0, 1 and 2 (G0, G1 and G2, respectively). As a result, a relaxivity hump between 10 and 100 MHz of the Larmor frequency progressively appeared with an increase in the size of the investigated contrast agent, indicative of a restricted rotational motion of the complexes as long as the size of the molecule increases. The same trend with a more pronounced effect was detectable in the presence of calcium. The relaxivity enhancement for the Ca2+ adducts, primarily caused by an increase of the hydration state of Gd3+, went from ca. 130\textpercent for the monomeric probe to ca. 310\textpercent for the G2 dendrimer conjugate at 0.5 T and 25 \mbox{$^\circ$}C. T1 weighted magnetic resonance images acquired at 1 T displayed the strong ability of these systems to change their contrast according to the presence of calcium at this field, thus laying the basis for promising future in vivo applications.}

{Dorsal human midbrain contains two nuclei with clear laminar organization, the superior and inferior colliculi. These nuclei extend in depth between the superficial dorsal surface of midbrain and a deep midbrain nucleus, the periaqueductal gray matter (PAG). The PAG, in turn, surrounds the cerebral aqueduct (CA). This study examined the use of two depth metrics to characterize depth and thickness relationships within dorsal midbrain using the superficial surface of midbrain and CA as references. The first utilized nearest-neighbor Euclidean distance from one reference surface, while the second used a level-set approach that combines signed distances from both reference surfaces. Both depth methods provided similar functional depth profiles generated by saccadic eye movements in a functional MRI task, confirming their efficacy for delineating depth for superficial functional activity. Next, the boundaries of the PAG were estimated using Euclidean distance together with elliptical fitting, indicating that the PAG can be readily characterized by a smooth surface surrounding PAG. Finally, we used the level-set approach to measure tissue depth between the superficial surface and the PAG, thus characterizing the variable thickness of the colliculi. Overall, this study demonstrates depth-mapping schemes for human midbrain that enables accurate segmentation of the PAG and consistent depth and thickness estimates of the superior and inferior colliculi.}

Spreading processes are conventionally monitored on a macroscopic level by counting the number of incidences over time. The spreading process can then be modeled either on the microscopic level, assuming an underlying interaction network, or directly on the macroscopic level, assuming that microscopic contributions are negligible. The macroscopic characteristics of both descriptions are commonly assumed to be identical. In this work we show that these characteristics of microscopic and macroscopic descriptions can be different due to coalescence, i.e., a node being activated at the same time by multiple sources. In particular, we consider a (microscopic) branching network (probabilistic cellular automaton) with annealed connectivity disorder, record the macroscopic activity, and then approximate this activity by a (macroscopic) branching process. In this framework we analytically calculate the effect of coalescence on the collective dynamics. We show that coalescence leads to a universal nonlinear scaling function for the conditional expectation value of successive network activity. This allows us to quantify the difference between the microscopic model parameter and established estimates of the macroscopic branching parameter. To overcome this difference, we propose a nonlinear estimator that correctly infers the microscopic model parameter for all system sizes.

{Purpose The purpose of this study is to introduce a novel design method of a shim coil array specifically optimized for whole brain shimming and to compare the performance of the resulting coils to conventional spherical harmonic shimming. Methods The proposed design approach is based on the stream function method and singular value decomposition. Eighty-four field maps from 12 volunteers measured in seven different head positions were used during the design process. The cross validation technique was applied to find an optimal number of coil elements in the array. Additional 42 field maps from 6 further volunteers were used for an independent validation. A bootstrapping technique was used to estimate the required population size to achieve a stable coil design. Results Shimming using 12 and 24 coil elements outperforms fourth- and fifth-order spherical harmonic shimming for all measured field maps, respectively. Coil elements show novel coil layouts compared to the conventional spherical harmonic coils and existing multi-coils. Both leave-one-out and independent validation demonstrate the generalization ability of the designed arrays. The bootstrapping analysis predicts that field maps from approximately 140 subjects need to be acquired to arrive at a stable design. Conclusions The results demonstrate the validity of the proposed method to design a shim coil array matched to the human brain anatomy, which naturally satisfies the laws of electrodynamics. The design method may also be applied to develop new shim coil arrays matched to other human organs.}

{Introduction: Patients with functional (psychogenic) movement disorders (FMD) can have various motor and non-motor symptoms including anxiety and depression. An aberrant integration of the sensorimotor and affective processing, and a dysfunctional sense of agency have been proposed as the key mechanisms involved in FMD development suggesting that FMD have features of a multiple network disorder. Therefore, exploration of brain connectivity mapping could provide important insights in the pathophysiology of FMD. The aim of this study was to analyze resting state fMRI to avoid variability related to aberrant movements or somatosensory misperception while taking into account anxiety and depression as confounding factors. To avoid selection bias, we used the eigen-vector centrality (EC) mapping to allow data-driven detection of connectivity hubs with rich connections to other regions of the brain that are highly connected. Methods: Forty-four patients with FMD (34F, age 45$\pm$(SD)9 years, disease duration 8.4$\pm$5,6 years) with heterogeneous motor phenotypes were compared to 44 matched control subjects (33F, 44$\pm$10y). The anxiety state (STAIX-1) and trait (STAIX-2) and Beck depression inventory (BDI-II) were recorded. All subjects were instructed to watch a cross for 10 minutes during rs-fMRI acquisition (3T, Siemens, Skyra: T2\textasteriskcentered-weighted gradient echo echo-planar imaging, TR\textequals2s, TE\textequals30ms, FA\textequals90, 300 scans). The preprocessing and statistics were performed with the SPM12 software (London, UK). General connectivity was derived from EC mapping (Lohmann et al., 2010) with a new correlation metric called ReLU correlation (Lohmann et al., 2018). For comparison, we also computed the EC with three other correlation metrics. Group analysis was based on a GLM implementing the contrast between both groups of subjects including \textquotesingleSTAIX-2\textquotesingle, \textquotesingleage\textquotesingle, \textquotesinglegender\textquotesingle, and \textquotesingleantidepressant drugs\textquotesingle as additional covariates. For selective connectivity, voxel-wise correlations using seeds derived from the results of the group EC analysis were computed in each subject. Group results were obtained with non-parametric statistics based on threshold-free cluster enhancement (TFCE toolbox, Jena, Germany) using the family-wise error correction at p\textless0.05. Results: FMD patients in comparison with controls showed decreased general connectivity in the SMA and midcingulate bilaterally and in the right superior frontal gyrus and right insula. This result became significant if the STAIX-2 was used as the nuisance covariate in the model (Figure 1). Selective connectivity from the seeds in the SMA and midcingulate showed disconnections in FMD patients in comparison with controls in multiple regions involving Rolandic areas, superior temporal gyri, posterior cingulate, insula, inferolateral prefrontal cortex, visual cortex and cerebellum (Figure 2). We obtained similar results with all correlation metrics, however, the RLC approach appeared to be most sensitive for the detection of EC differences between both groups. Conclusions: Despites motor symptoms variability, the brain of FMD patients exhibited a uniform connectivity pattern involving the SMA and midcingulate, which were generally less connected with all other cortical and subcortical regions than in controls. The results are in line with previous reports (Roelofs et al., 2019) using different approaches to rs-fMRI or brain morphometry analyses and fit well to our knowledge of the functional role of these regions and their possible malfunctions. Disconnected SMA is potentially associated with dysfunctional planning of voluntary movements or with insufficient suppression of unwanted actions (Nachev et al., 2008; Ruan et al., 2018). Disconnection of the midcingulate may affect multisensory orientation, nociceptive processing or nocifensive behavior (Vogt, 2016; Ospina et al., 2019). Our findings thus support FMD as a network disorder affecting hubs connecting motor and non-motor networks. Supported by the grant AZV 16-29651A.}

{Individual characteristics in human decision-making are often quantified by fitting a parametric cognitive model to subjects\textquotesingle behavior and then studying differences between them in the associated parameter space. However, these models often fit behavior more poorly than recurrent neural networks (RNNs), which are more flexible and make fewer assumptions about the underlying decision-making processes. Unfortunately, the parameter and latent activity spaces of RNNs are generally high-dimensional and uninterpretable, making it hard to use them to study individual differences. Here, we show how to benefit from the flexibility of RNNs while representing individual differences in a low-dimensional and interpretable space. To achieve this, we propose a novel end-to-end learning framework in which an encoder is trained to map the behavior of subjects into a low-dimensional latent space. These low-dimensional representations are used to generate the parameters of individual RNNs corresponding to the decision-making process of each subject. We introduce terms into the loss function that ensure that the latent dimensions are informative and disentangled, i.e., encouraged to have distinct effects on behavior. This allows them to align with separate facets of individual differences. We illustrate the performance of our framework on synthetic data as well as a dataset including the behavior of patients with psychiatric disorders.}

{Individuals differ in how they learn from experience. In Pavlovian conditioning models, where cues predict reinforcer delivery at a different goal location, some animals-called sign-trackers-come to approach the cue, whereas others, called goal-trackers, approach the goal. In sign-trackers, model-free phasic dopaminergic reward-prediction errors underlie learning, which renders stimuli \textquotesinglewanted\textquotesingle. Goal-trackers do not rely on dopamine for learning and are thought to use model-based learning. We demonstrate this double dissociation in 129 male humans using eye-tracking, pupillometry and functional magnetic resonance imaging informed by computational models of sign- and goal-tracking. We show that sign-trackers exhibit a neural reward prediction error signal that is not detectable in goal-trackers. Model-free value only guides gaze and pupil dilation in sign-trackers. Goal-trackers instead exhibit a stronger model-based neural state prediction error signal. This model-based construct determines gaze and pupil dilation more in goal-trackers.}

{The noradrenergic locus coeruleus (LC) is crucial for controlling brain and behavioral states. While synchronous stimulation of LC neurons evokes a single activated cortical state with increased high-frequency power, little is known about how spontaneous patterns of LC population activity drive cortical states. Since LC neurons selectively project to specific forebrain regions, we hypothesized that individual LC ensembles produce different cortical states. We recorded up to 34 single units simultaneously in the rat LC and used non-negative matrix factorization to identify spontaneously activated ensembles of co-active LC neurons. The ensembles were active mostly at different times and were simultaneously active only rarely. We assessed cortical state in area 24a by examining local field potential power spectrograms triggered on activations of individual LC ensembles. We observed four spectrotemporally-distinct cortical states associated with activation of specific LC ensembles. Thus, distinct spontaneously active LC ensembles contribute to unexpectedly diverse cortical states.}

{When knowledge is scarce, it is adaptive to seek further information to resolve uncertainty and obtain a more accurate worldview. Biases in such information-seeking behavior can contribute to the maintenance of inaccurate views. Here, we investigate whether predispositions for uncertainty-guided information seeking relate to individual differences in dogmatism, a phenomenon linked to entrenched beliefs in political, scientific, and religious discourse. We addressed this question in a perceptual decision-making task, allowing us to rule out motivational factors and isolate the role of uncertainty. In two independent general population samples (n \textequals 370 and n \textequals 364), we show that more dogmatic participants are less likely to seek out new information to refine an initial perceptual decision, leading to a reduction in overall belief accuracy despite similar initial decision performance. Trial-by-trial modeling revealed that dogmatic participants placed less reliance on internal signals of uncertainty (confidence) to guide information search, rendering them less likely to seek additional information to update beliefs derived from weak or uncertain initial evidence. Together, our results highlight a cognitive mechanism that may contribute to the formation of dogmatic worldviews.}

{PURPOSE: To develop a robust design of a human head double-tuned 31 P/1 H array, which provides good performance at both 31 P and 1 H frequencies for MR spectroscopic imaging at 9.4T. METHODS: Increasing the number of surface loops in a human head array improves the peripheral signal-to-noise ratio (SNR), while the central SNR doesn\textquotesinglet substantially change. High peripheral SNR can contaminate MR spectroscopic imaging data at both 1 H and 31 P frequency. To minimize this effect, we limited the number of elements in the 31 P array to 10, i.e., 8 transceiver surface loops circumscribing the head and 2 receive "vertical" loops placed at the superior location. The 1 H-portion of the array also consists of 10 elements, i.e., 8 transceiver surface loops circumscribing the head and 2 transceiver "vertical" loops at the superior location of the head. Both the 31 P array and 1 H array are placed in a single layer at the same distance to the head, which provides high loading and, thus, a good performance for both arrays. RESULTS: Transmit efficiency of the 1 H-portion of the double-tuned array was very similar to that of the single-tuned arrays of similar size. Also, addition of the cross-loops substantially improved the brain coverage. CONCLUSION: We developed a novel 31 P/1 H double-tuned array for MR spectroscopic imaging of a human brain at 9.4T. Placing both 31 P and 1 H loops in a single layer provides for high transmit efficiency at both frequencies without compromising SNR near the brain center at the 31 P-frequency. Addition of the cross-loops at the superior location improves the brain coverage.}