Automatic Cerebral Hemisphere Segmentation in Rat MRI with Ischemic Lesions via Attention-based Convolutional Neural Networks.

We present MedicDeepLabv3+, a convolutional neural network that is the first completely automatic method to segment cerebral hemispheres in magnetic resonance (MR) volumes of rats with ischemic lesions. MedicDeepLabv3+ improves the state-of-the-art DeepLabv3+ with an advanced decoder, incorporating spatial attention layers and additional skip connections that, as we show in our experiments, lead to more precise segmentations. MedicDeepLabv3+ requires no MR image preprocessing, such as bias-field correction or registration to a template, produces segmentations in less than a second, and its GPU memory requirements can be adjusted based on the available resources. We optimized MedicDeepLabv3+ and six other state-of-the-art convolutional neural networks (DeepLabv3+, UNet, HighRes3DNet, V-Net, VoxResNet, Demon) on a heterogeneous training set comprised by MR volumes from 11 cohorts acquired at different lesion stages. Then, we evaluated the trained models and two approaches specifically designed for rodent MRI skull stripping (RATS and RBET) on a large dataset of 655 MR rat brain volumes. In our experiments, MedicDeepLabv3+ outperformed the other methods, yielding an average Dice coefficient of 0.952 and 0.944 in the brain and contralateral hemisphere regions. Additionally, we show that despite limiting the GPU memory and the training data, our MedicDeepLabv3+ also provided satisfactory segmentations. In conclusion, our method, publicly available at https://github.com/jmlipman/MedicDeepLabv3Plus , yielded excellent results in multiple scenarios, demonstrating its capability to reduce human workload in rat neuroimaging studies.

Corrigendum: RAFF-4, Magnetization Transfer and Diffusion Tensor MRI of Lysophosphatidylcholine Induced Demyelination and Remyelination in Rats.

[This corrects the article DOI: 10.3389/fnins.2021.625167.].

RAFF-4, Magnetization Transfer and Diffusion Tensor MRI of Lysophosphatidylcholine Induced Demyelination and Remyelination in Rats.

Remyelination is a naturally occurring response to demyelination and has a central role in the pathophysiology of multiple sclerosis and traumatic brain injury. Recently we demonstrated that a novel MRI technique entitled Relaxation Along a Fictitious Field (RAFF) in the rotating frame of rank n (RAFFn) achieved exceptional sensitivity in detecting the demyelination processes induced by lysophosphatidylcholine (LPC) in rat brain. In the present work, our aim was to test whether RAFF4, along with magnetization transfer (MT) and diffusion tensor imaging (DTI), would be capable of detecting the changes in the myelin content and microstructure caused by modifications of myelin sheets around axons or by gliosis during the remyelination phase after LPC-induced demyelination in the corpus callosum of rats. We collected MRI data with RAFF4, MT and DTI at 3 days after injection (demyelination stage) and at 38 days after injection (remyelination stage) of LPC (n = 12) or vehicle (n = 9). Cell density and myelin content were assessed by histology. All MRI metrics detected differences between LPC-injected and control groups of animals in the demyelination stage, on day 3. In the remyelination phase (day 38), RAFF4, MT parameters, fractional anisotropy, and axial diffusivity detected signs of a partial recovery consistent with the remyelination evident in histology. Radial diffusivity had undergone a further increase from day 3 to 38 and mean diffusivity revealed a complete recovery correlating with the histological assessment of cell density attributed to gliosis. The combination of RAFF4, MT and DTI has the potential to differentiate between normal, demyelinated and remyelinated axons and gliosis and thus it may be able to provide a more detailed assessment of white matter pathologies in several neurological diseases.

Automated joint skull-stripping and segmentation with Multi-Task U-Net in large mouse brain MRI databases.

Skull-stripping and region segmentation are fundamental steps in preclinical magnetic resonance imaging (MRI) studies, and these common procedures are usually performed manually. We present Multi-task U-Net (MU-Net), a convolutional neural network designed to accomplish both tasks simultaneously. MU-Net achieved higher segmentation accuracy than state-of-the-art multi-atlas segmentation methods with an inference time of 0.35 s and no pre-processing requirements. We trained and validated MU-Net on 128 T2-weighted mouse MRI volumes as well as on the publicly available MRM NeAT dataset of 10 MRI volumes. We tested MU-Net with an unusually large dataset combining several independent studies consisting of 1782 mouse brain MRI volumes of both healthy and Huntington animals, and measured average Dice scores of 0.906 (striati), 0.937 (cortex), and 0.978 (brain mask). Further, we explored the effectiveness of our network in the presence of different architectural features, including skip connections and recently proposed framing connections, and the effects of the age range of the training set animals. These high evaluation scores demonstrate that MU-Net is a powerful tool for segmentation and skull-stripping, decreasing inter and intra-rater variability of manual segmentation. The MU-Net code and the trained model are publicly available at https://github.com/Hierakonpolis/MU-Net.

Cognitive disturbances in the cuprizone model of multiple sclerosis.

Cognitive problems frequently accompany neurological manifestations of multiple sclerosis (MS). However, during screening of preclinical candidates, assessments of behaviour in mouse models of MS typically focus on locomotor activity. In the present study, we analysed cognitive behaviour of 9 to 10-week-old female C57Bl/6J mice orally administered with the toxin cuprizone that induces demyelination, a characteristic feature of MS. Animals received 400 mg/kg cuprizone daily for 2 or 4 weeks, and their performance was compared with that of vehicle-treated mice. Cuprizone-treated animals showed multiple deficits in short touchscreen-based operant tasks: they responded more slowly to visual stimuli, rewards and made more errors in a simple rule-learning task. In contextual/cued fear conditioning experiments, cuprizone-treated mice showed significantly lower levels of contextual freezing than vehicle-treated mice. Diffusion tensor imaging showed treatment-dependent changes in fractional anisotropy as well as in axial and mean diffusivities in different white matter areas. Lower values of fractional anisotropy and axial diffusivity in cuprizone-treated mice indicated developing demyelination and/or axonal damage. Several diffusion tensor imaging measurements correlated with learning parameters. Our results show that translational touchscreen operant tests and fear conditioning paradigms can reliably detect cognitive consequences of cuprizone treatment. The suggested experimental approach enables screening novel MS drug candidates in longitudinal experiments for their ability to improve pathological changes in brain structure and reverse cognitive deficits.

RatLesNetv2: A Fully Convolutional Network for Rodent Brain Lesion Segmentation.

We present a fully convolutional neural network (ConvNet), named RatLesNetv2, for segmenting lesions in rodent magnetic resonance (MR) brain images. RatLesNetv2 architecture resembles an autoencoder and it incorporates residual blocks that facilitate its optimization. RatLesNetv2 is trained end to end on three-dimensional images and it requires no preprocessing. We evaluated RatLesNetv2 on an exceptionally large dataset composed of 916 T2-weighted rat brain MRI scans of 671 rats at nine different lesion stages that were used to study focal cerebral ischemia for drug development. In addition, we compared its performance with three other ConvNets specifically designed for medical image segmentation. RatLesNetv2 obtained similar to higher Dice coefficient values than the other ConvNets and it produced much more realistic and compact segmentations with notably fewer holes and lower Hausdorff distance. The Dice scores of RatLesNetv2 segmentations also exceeded inter-rater agreement of manual segmentations. In conclusion, RatLesNetv2 could be used for automated lesion segmentation, reducing human workload and improving reproducibility. RatLesNetv2 is publicly available at https://github.com/jmlipman/RatLesNetv2.

Spontaneous BOLD waves - A novel hemodynamic activity in Sprague-Dawley rat brain detected by functional magnetic resonance imaging.

We report spontaneous hemodynamic activity termed "Spontaneous BOLD Waves" (SBWs) detected by BOLD fMRI in Sprague-Dawley rats under medetomidine anesthesia. These SBWs, which lasted several minutes, were observed in cortex, thalamus and hippocampus. The SBWs' correlates were undetectable in electrophysiological recordings, suggesting an exclusive gliovascular phenomenon dissociated from neuronal activity. SBWs were insensitive to the NMDA receptors antagonist MK-801 but were inhibited by the α1-adrenoceptor blocker prazosin. Since medetomidine is a potent agonist of α2 adrenoceptors, we suggested that imbalance in α1/α2 receptor-mediated signalling pathways alter the vascular reactivity leading to SBWs. The frequency of SBWs increased with intensity of mechanical lung ventilation despite the stable pH levels. In summary, we present a novel type of propagating vascular brain activity without easily detectable underlying neuronal activity, which can be utilized to study the mechanisms of vascular reactivity in functional and pharmacological MRI and has practical implications for designing fMRI experiments in anesthetized animals.

Awake Rat Brain Functional Magnetic Resonance Imaging Using Standard Radio Frequency Coils and a 3D Printed Restraint Kit.

Functional magnetic resonance imaging (fMRI) is a powerful noninvasive tool for studying spontaneous resting state functional connectivity (RSFC) in laboratory animals. Brain function can be significantly affected by generally used anesthetics, however, rendering the need for awake imaging. Only a few different awake animal habituation protocols have been presented, and there is a critical need for practical and improved low-stress techniques. Here we demonstrate a novel restraint approach for awake rat RSFC studies. Our custom-made 3D printed restraint kit is compatible with a standard Bruker Biospin MRI rat bed, rat brain receiver coil, and volume transmitter coil. We also implemented a progressive habituation protocol aiming to minimize the stress experienced by the rats, and compared RSFC between awake, lightly sedated, and isoflurane-anesthetized rats. Our results demonstrated that the 3D printed restraint kit was suitable for RSFC studies of awake rats. During the short 4-day habituation period, the plasma corticosterone concentration, movement, and heart rate, which were measured as stress indicators, decreased significantly, indicating adaptation to the restraint protocol. Additionally, 10 days after the awake MRI session, rats exhibited no signs of depression or anxiety based on open-field and sucrose preference behavioral tests. The RSFC data revealed significant changes in the thalamo-cortical and cortico-cortical networks between the awake, lightly sedated, and anesthetized groups, emphasizing the need for awake imaging. The present work demonstrates the feasibility of our custom-made 3D printed restraint kit. Using this kit, we found that isoflurane markedly affected brain connectivity compared with that in awake rats, and that the effect was less pronounced, but still significant, when light isoflurane sedation was used instead.

Detection of Hyperexcitability by Functional Magnetic Resonance Imaging after Experimental Traumatic Brain Injury.

Diagnosis of ongoing epileptogenesis and associated hyperexcitability after brain injury is a major challenge. Given that increased neuronal activity in the brain triggers a blood oxygenation level-dependent (BOLD) response in functional magnetic resonance imaging (fMRI), we hypothesized that fMRI could be used to identify the brain area(s) with hyperexcitability during post-injury epileptogenesis. We applied fMRI to detect onset and spread of BOLD activation after pentylenetetrazol (PTZ)-induced seizures (PTZ, 30 mg/kg, intraperitoneally) in 16 adult male rats at 2 months after lateral fluid percussion (FPI)-induced traumatic brain injury (TBI). In sham-operated controls, onset of the PTZ-induced BOLD response was bilateral and first appeared in the cortex. After TBI, 5 of 9 (56%) rats exhibited ipsilateral perilesional cortical BOLD activation, followed by activation of the contralateral cortex. In 4 of 9 (44%) rats, onset of BOLD response was bilateral. Interestingly, latency from the PTZ injection to onset of the BOLD response increased in the following order: sham-operated controls (ipsilateral 132 ± 57 sec, contralateral 132 ± 57 sec; p > 0.05) < TBI with bilateral BOLD onset (ipsilateral 176 ± 54 sec, contralateral 178 ± 52 sec; p > 0.05) < TBI with ipsilateral BOLD onset (ipsilateral 406 ± 178 sec, contralateral 509 ± 140 sec; p < 0.05). Cortical lesion area did not differ between rats with ipsilateral versus bilateral BOLD onset (p > 0.05). In the group of rats with ipsilateral onset of PTZ-induced BOLD activation, none of the rats showed a robust bilateral thalamic BOLD response, only 1 of 5 rats had robust ipsilateral thalamic calcifications, and 4 of 5 rats had perilesional astrocytosis. These findings suggest the evolution of the epileptogenic zone in the perilesional cortex after TBI, which is sensitive to PTZ-induced hyperexcitability. Further studies are warranted to explore the evolution of thalamo-cortical pathology as a driver of epileptogenesis after lateral FPI.

Orientation selective deep brain stimulation.

OBJECTIVE: Target selectivity of deep brain stimulation (DBS) therapy is critical, as the precise locus and pattern of the stimulation dictates the degree to which desired treatment responses are achieved and adverse side effects are avoided. There is a clear clinical need to improve DBS technology beyond currently available stimulation steering and shaping approaches. We introduce orientation selective neural stimulation as a concept to increase the specificity of target selection in DBS. APPROACH: This concept, which involves orienting the electric field along an axonal pathway, was tested in the corpus callosum of the rat brain by freely controlling the direction of the electric field on a plane using a three-electrode bundle, and monitoring the response of the neurons using functional magnetic resonance imaging (fMRI). Computational models were developed to further analyze axonal excitability for varied electric field orientation. MAIN RESULTS: Our results demonstrated that the strongest fMRI response was observed when the electric field was oriented parallel to the axons, while almost no response was detected with the perpendicular orientation of the electric field relative to the primary fiber tract. These results were confirmed by computational models of the experimental paradigm quantifying the activation of radially distributed axons while varying the primary direction of the electric field. SIGNIFICANCE: The described strategies identify a new course for selective neuromodulation paradigms in DBS based on axonal fiber orientation.

Implantable RF-coil with multiple electrodes for long-term EEG-fMRI monitoring in rodents.

BACKGROUND: Simultaneous EEG-fMRI is a valuable tool in the clinic as it provides excellent temporal and spatial information about normal and diseased brain function. In pre-clinical research with small rodents, obtaining simultaneous EEG-fMRI in longitudinal studies faces a number of challenges, including issues related to magnetic susceptibility artifacts. NEW METHOD: Here, we demonstrate a method for permanent MRI RF-coil and EEG electrode implantation in rats that is suitable for long-term chronic follow-up studies in both stimulus and resting-state fMRI paradigms. RESULTS: Our findings showed that the screw-free implantation method is well suited for long-term follow-up studies in both freely moving video-EEG settings and fMRI without causing MRI susceptibility artifacts. Furthermore, the results demonstrated that a multimodal approach can be used to track the progression of structural and functional changes. COMPARISON WITH EXISTING METHODS: The quality of both MRI and EEG data were comparable to those obtained with traditional methods with the benefit of combining them into artifact-free simultaneous recordings. The signal-to-noise ratios of the MRI images obtained with the implanted RF-coil were similar to those using a quadrature coil and were therefore suitable for resting-state fMRI experiments. Similarly, EEG data collected with the RF-coil/electrode set-up were comparable to EEG recorded with traditional epidural screw electrodes. CONCLUSION: This new multimodal EEG-fMRI approach provides a novel tool for concomitant analysis and follow-up of anatomic and functional MRI, as well as electrographic changes in a preclinical research.

Global Functional Connectivity Differences between Sleep-Like States in Urethane Anesthetized Rats Measured by fMRI.

Sleep is essential for nervous system functioning and sleep disorders are associated with several neurodegenerative diseases. However, the macroscale connectivity changes in brain networking during different sleep states are poorly understood. One of the hindering factors is the difficulty to combine functional connectivity investigation methods with spontaneously sleeping animals, which prevents the use of numerous preclinical animal models. Recent studies, however, have implicated that urethane anesthesia can uniquely induce different sleep-like brain states, resembling rapid eye movement (REM) and non-REM (NREM) sleep, in rodents. Therefore, the aim of this study was to assess changes in global connectivity and topology between sleep-like states in urethane anesthetized rats, using blood oxygenation level dependent (BOLD) functional magnetic resonance imaging. We detected significant changes in corticocortical (increased in NREM-like state) and corticothalamic connectivity (increased in REM-like state). Additionally, in graph analysis the modularity, the measure of functional integration in the brain, was higher in NREM-like state than in REM-like state, indicating a decrease in arousal level, as in normal sleep. The fMRI findings were supported by the supplementary electrophysiological measurements. Taken together, our results show that macroscale functional connectivity changes between sleep states can be detected robustly with resting-state fMRI in urethane anesthetized rats. Our findings pave the way for studies in animal models of neurodegenerative diseases where sleep abnormalities are often one of the first markers for the disorder development.

Comparison of seven different anesthesia protocols for nicotine pharmacologic magnetic resonance imaging in rat.

Pharmacologic MRI (phMRI) is a non-invasive in vivo imaging method, which can evaluate the drug effects on the brain and provide complementary information to ex vivo techniques. The preclinical phMRI studies usually require anesthesia to reduce the motion and stress of the animals. The anesthesia, however, is a crucial part of the experimental design, as it may modulate the neural drug-induced (de)activation and hemodynamic coupling. Therefore, the aim of the present study was to address this methodologic question by performing phMRI experiments with five anesthetics (α-chloralose, isoflurane, medetomidine, thiobutabarbital, and urethane) and seven anesthesia protocols. Nicotine, a widely studied psychostimulant, was administered to rats while measuring blood oxygenation level-dependent (BOLD) signals. Notably different responses were observed depending on the anesthetic used. The highest responses were measured in urethane-anesthetized rats whereas the responses were hardly noticeable in α-chloralose group. As urethane is not commonly used in phMRI, hemodynamic coupling under urethane anesthesia was investigated with functional cerebral blood flow (CBF) and volume-weighted (CBVw) imaging, and simultaneous electrophysiologic and BOLD measurements. The BOLD, CBF, and CBVw measurements in response to nicotine were highly correlated (R(2) ≥ 0.70, p<0.001). BOLD values correlated well (R(2)=0.43, p<10(-6)) with local field potential (LFP) spectral power (13-70Hz) during pharmacologic stimulation. These findings indicate that urethane anesthesia combined with BOLD contrast provides a robust protocol for nicotine phMRI studies. As urethane has mild effects to individual receptor systems, and coupling between electrophysiologic activity and hemodynamic response is maintained, this anesthetic may also be suitable for other phMRI studies.

Involvement of NMDA receptor subtypes in cortical spreading depression in rats assessed by fMRI.

Cortical spreading depression (CSD) is a phenomenon implicated in migraine with aura and associated with other neurological disorders (e.g. stroke, brain trauma). Current evidence points to the essential role of NMDA receptors in CSD mechanisms. However, the roles of multiple subunits of NMDA receptors expressed in neurons, glia and blood vessels in vivo, are little explored. Using BOLD fMRI of urethane anesthetized rats as an integrative CSD readout, we tested the involvement of different NMDA receptor subtypes in CSD induction and propagation. Rats were treated with a non-selective NMDA blocker (MK-801), NR2B antagonist (ifenprodil) or a NR2A selective antagonist (TCN-201). CSD was induced during fMRI scanning by application of KCl onto the cerebral cortex and fMRI data were collected by 9.4 T MRI. The non-specific NMDA antagonist MK-801 completely blocked CSD, which was not observed in the NR2A group where TCN-201 did not alter the CSD features. Unexpectedly, the NR2B specific antagonist ifenprodil largely promoted the initial negative phase of the BOLD CSD response, likely due to altered neurovascular coupling. Our data suggest key roles and differential involvement of NMDA receptor subtypes in CSD generation and propagation, highlighting an important role for the NR2B subtype.

Fast vascular component of cortical spreading depression revealed in rats by blood pulsation imaging.

Cortical spreading depression (CSD) is a slowly propagating wave of depolarization of neurons and glia and has a less characterized vascular component. CSD is a commonly used phenomenon to test new methods of live brain imaging. Application of a blood pulsations imaging (BPI) technique to study of CSD induced with high-potassium solution in rat cortex allowed us to visualize for the first time the novel vascular component of a CSD wave. In our study, this wave component propagated in the limited part of the cortex along the bow-shaped trajectory in sharp contrast with concentric development of CSD measured by concurrently applied optical intrinsic signal (OIS) imaging technique. It was associated with a significant increase of the blood pulsations amplitude (BPA), started with a delay of 20 to 90 s comparing to signal measured with OIS, and propagated 40% faster than OIS signal. These findings suggest that the BPA and slower change of the cerebral blood volume are not directly related to each other even though both characterize the same vascular system. Our study indicates that the BPI technique could be used for characterization of the new pulsatile vascular component of CSDs in animal models of migraine, stroke, and brain trauma.

Opposite reactivity of meningeal versus cortical microvessels to the nitric oxide donor glyceryl trinitrate evaluated in vivo with two-photon imaging.

Vascular changes underlying headache in migraine patients induced by Glyceryl trinitrate (GTN) were previously studied with various imaging techniques. Despite the long history of medical and experimental use of GTN, its effects on the brain vasculature are still poorly understood presumably due to low spatial resolution of the imaging modalities used so far. We took advantage of the micrometer-scale vertical resolution of two-photon microscopy to differentiate between the vasodynamic effects of GTN on meningeal versus cortical vessels imaged simultaneously in anesthetized rats through either thinned skull or glass-sealed cranial window. Intermediate and small calibre vessels were visualized in vivo by imaging intravascular fluorescent dextran, and detection of blood flow direction allowed identification of individual arterioles and venules. We found that i.p.-injected GTN induced a transient constriction of meningeal arterioles, while their cortical counterparts were, in contrast, dilated. These opposing effects of GTN were restricted to arterioles, whereas the effects on venules were insignificant. Interestingly, the NO synthase inhibitor L-NAME did not affect the diameter of meningeal vessels but induced a constriction of cortical vessels. The different cellular environment in cortex versus meninges as well as distinct vessel wall anatomical features probably play crucial role in the observed phenomena. These findings highlight differential region- and vessel-type-specific effects of GTN on cranial vessels, and may implicate new vascular mechanisms of NO-mediated primary headaches.

Parenchymal spin-lock fMRI signals associated with cortical spreading depression.

We found novel types of parenchymal functional magnetic resonance imaging (fMRI) signals in the rat brain during large increases in metabolism. Cortical spreading depression (CSD), a self-propagating wave of cellular activation, is associated with several pathologic conditions such as migraine and stroke. It was used as a paradigm to evoke transient neuronal depolarization leading to enhanced energy consumption. Activation of CSD was investigated using spin-lock (SL), diffusion, blood oxygenation level-dependent and cerebral blood volume fMRI techniques. Our results show that the SL-fMRI signal is generated by endogenous parenchymal mechanisms during CSD propagation, and these mechanisms are not associated with hemodynamic changes or cellular swelling. Protein phantoms suggest that pH change alone does not explain the observed SL-fMRI signal changes. However, increased amounts of inorganic phosphates released from high-energy phosphates combined with pH changes may produce SL- power-dependent longitudinal relaxation in the rotating frame (R1ρ) changes in protein phantoms that are similar to those observed during CSD, as seen before in acute ischemia under our experimental conditions. This links SL-fMRI changes intimately to energy metabolism and supports the use of the SL technique as a new, promising functional approach for noninvasive imaging of metabolic transitions in the active or pathologic brain.