3. Example “vascular connectome” and specific vascular pathways connecting pairs of brain regions. For both (A) Freesurfer’s Destrieux 2009 GM parcellation and (B) Recobundle White-Matter bundle atlas, the connectome obtained for one subject is shown. For each, two example region-to-region connections were selected from the tracts computed previously to illustrate the complex pathway connecting those regions.

2. Vascular tractography. (A) shows 10k tracts reconstructed from the 200k computed, with two cross-sections highlighted in yellow and purple. In yellow, the tracts and ‘vesselness’ scores are overlaid by the ODF at crossings, while the ri­ghtmost shows the main directions (tensor) of the ODFs. (B) In the purple cross-section, the tracts and vessels are overlaid with transparency to demonstrate the correspondence.

Figure 2: (A) Connectivity matrix (i.e. mean connectivity values, n = 20) and (B) circular connectome (–log₁₀(p-value)) of 15 brainstem nuclei (red bracket). For both (A) and (B), for display purposes, we thresholded the results at p<0.0005 Bonferroni-corrected. C) On top, bar plot displaying laterality index for bilateral seeds only. On bottom, bar plot displaying the connectivity degree of each seed: the horizontal line refers to the average degree. Seeds exceeding the average (i.e. IC, VTA-PBP, LPB, MPB, Ve, sMRt, VSM, iMRT) were considered hubs and marked with a red star.

Figure 3: Functional connectome of (A) left iMRt and (B) left MPB (p<0.0005 Bonferroni corrected for display purposes). (A) We confirmed iMRt connectivity to: ION, previously found in cats¹²; Ve¹³; VSM¹⁴ for taste-elicited ingestion/rejection responses; PAG for pain modulation¹⁵. Expected connectivity with SC and LPB were low¹¹. (B) MPB showed expected connectivity to DR, thalamus, and (weaker) VSM and amygdala, yet we did not observe connectivity to RMg. Connectivity to insular and cingulate cortex was strong¹¹. Connectivity to PAG and VTA was in line with rat studies^16,17.

Figure 3. (A) The average of fMRI time segments centering at sensory dominant co-activations (N=343). (B) The EEG alpha power, HR, RV changes in the same group of time segments. (C) Left: The alpha-rsfMRI correlation maps before and after regressing out the fMRI sequence in (A). Right: the lag-dependent alpha-rsfMRI correlation within the three ROIs before and after regressing out the fMRI sequence. The shaded regions represent area within 1 S.E.M. (D-E): The physio-rsfMRI correlation maps before and after regressing out the fMRI sequence.

Figure 2. RsfMRI correlations with physiological signals show similar patterns as the alpha-rsfMRI correlations. (A) The z-scored correlation maps of RV/HR with rsfMRI signals at various time lags are similar to the alpha-rsfMRI correlation maps. (B) Cross correlation function between the alpha-rsfMRI correlation maps and the RV/HR-rsfMRI correlation maps.

Fig. 2 Predicted and actual activation maps of 3 representative subjects in the WM (2BK), social (RANDOM), emotion (FACES) behavioral domains.

Fig. 1 Modeling pipeline. Group independent component analysis was implemented to parcellate data of all subjects into 50 shared brain networks, individuals’ connectivity maps for each network were derived following the dual regression procedure and were entered as features of the GCN. A customized brain parcellation of 399 ROIs were applied as graph nodes, individual FC matrix of those ROIs was calculated as the edges. Four Chebyshev spectral graph convolutional layers were used, instance normalization was added for each of the first three layers to reduce over-fitting.

Figure 1: Dictionary learning statistical map of resting-state large-scale network in four primate species using 7 components. Two high order network are illustrated for each species: Human (A): default mode and fronto-parietal control. Macaque (B), Mouse lemur (C), Marmoset (D): fronto-parietal and fronto-temporal.

Figure 2: Large-scale network functional atlas of the primate brains. Seven components of the dictionary learning analysis were concatenated and labeled based on their anatomical features. Human functional atlas (A) Macaque (B) Marmoset (C) Mouse lemur (D). Cerebral clusters were spatially separate (colored dashed line) and further used to extract their correlation strength with other clusters of interest.

Fig2.Cross-species CAP topography reveals evolutionarily conserved rsfMRI network engagements. Co-activation patterns (CAPs) from mice, macaques, and humans (p<0.05,Bonferroni corrected). Red indicates significant co-activation, blue significant co-deactivations. Abbreviations: Ctx-Cortex; SMN-Sensory-motor Network; DMN-Default-Mode Network; TH-Thalamus; Cd/Pu-Caudate/Putamen;VIS-Visual Network LN-Limbic Network;HCP-Hippocampus; ECN-Executive-Control Network;DAN-Dorsal-Attention Network; SN-Salience Network.

Fig3. CAPs exhibit infraslow fluctuations and gradual assembly/disassembly. A), C), E) Power-spectral density (PSD) of CAP-to-frame correlation timecourses (mean +/-SEM). B),D),F) CAP temporal assembly by time-lock averaging the highest peaks in the CAP timecourse.

Figure 1. Simultaneous recording of EGG and brain fMRI reveals distinct fMRI maps at different gastric conditions. (A) layouts of the 32-channel electrode array for EGG recording and plotted an example of EGG time series. (B) shows the power fluctuations of EGG at different frequencies. (CDE) are averaged fMRI maps during bradygastria, tachygastria, and rhythmic condition, together with examples of the power-frequency relationship curve. In fMRI maps, color encodes z-value that obtained by the distribution of fMRI intensity from the whole brain.

Figure 2. The EGG power network of the rhythmic EGG. (A) we assume that EGG power is extrinsic stimuli to the BOLD activity. We can estimate the response function H based on EGG power and the BOLD activity. According to the robustness of H, we can map the corresponding EGG power network. (B) shows the EGG power network at 5CPM. Color encodes t-score. The EGG power covers the nucleus of the solitary tract (NTS), spinal vestibular nucleus (SpVe), thalamic nucleus (VM), amygdala (APir), somatosensory cortex (S1BF), insular cortex (AIP), striatum (Cpu), orbital cortex (VO/LO.

Figure 1. Electrophysiology setup and correlations between pupil dynamics and LFP delta power from the ACC or LH. a) the experimental setup for electrophysiology and pupil detection. b) examples of positive and negative correlations between pupil diameter and LFP delta (1-4Hz) power changes from the ACC and LH during 15-min resting state trials. c) Distribution of correlation coefficient between pupil dynamics and LFP delta band power at the LH or ACC. d) Categorization of trials based on a positive or negative correlation between pupil dynamics and LH delta power.

Figure 3. Correlations between fMRI and ACC 1-4Hz calcium power in the whole brain. a) multimodal resting-state fMRI setup simultaneously recorded with optical fiber calcium signals and pupil dynamics. b) Distribution of correlation coefficient between pupil dynamics and ACC 1-4Hz calcium power and their examples of positive and negative correlations during 15-min trials. c) t-value maps of the correlation between fMRI and ACC 1-4Hz calcium power when the pupil dynamics and ACC 1-4Hz calcium power show negative correlations and location of the LH, indicated with white arrows.

Bilaterally synchronized cortex-specific oscillations in fMRI signal detected at 0.20-0.24Hz. A – A mask is defined by considering all voxels exhibiting power 5 standard deviations above the mean level detected in a dead rat. B – The fMRI signal in the mask voxels is band-pass filtered between 0.20 and 0.24 Hz and plotted over time, revealing ultra-slow amplitude modulations (here one representative scan). C – Snap shots of the band-pass filtered fMRI signal revealing bilateral synchronization.

Slow oscillations decrease in power with anesthesia. A – fMRI Spectral power maps (averaged across scans) reveal resonance at specific frequencies in different brain and body structures that decrease significantly in power under deep anesthesia. B – Comparison of the power in each frequency bin between conditions (6 scans per condition), normalized by the average power in the dead rat scans. C – Statistically significant differences are only considered if p<0.00083 (Bonferroni threshold).

Brain-wide BOLD responses upon optogenetic stimulation of ssTRN at different frequencies. (A) Illustration of atlas-based ROI definitions in the sensorimotor cortical, thalamic and midbrain regions, higher-order cortical and limbic regions, and basal ganglia regions. (B) BOLD activation maps at different frequencies of ssTRN stimulation (Bonferroni-corrected p<0.001). 1Hz stimulation evoked positive bilateral cortical and vCPu responses while 4-40 Hz stimulations evoked negative ipsilateral vCPu response. (C) BOLD profiles extracted from atlas-based ROIs.

Histological confirmation of ChR2 expression in inhibitory TRN neurons, optogenetic fMRI experiment setup and stimulation paradigm. (A) Confocal images of ChR2-mCherry expression in TRN with lower (left) and higher (right) magnification. Overlay of images reveals colocalization of the nuclear marker DAPI, inhibitory marker GABA, and ChR2-mCherry in the cell body of inhibitory TRN neurons (indicated by white arrows). (B) Optogenetic stimulation setup illustration (left) and a typical fMRI scan timeline with different optogenetic stimulation paradigms (right).

Fig 4. Shows the Visual Occipital regions’ normalized Z values for a specific subject for (a) mean tSNR = 74.2 and (b) mean tSNR = 17.8 which causes the similarity index calculated between the normative database and volunteer’s data to drop from 73% in (a) to 62% in (b) within a 95% confidence interval.

Fig 3. Shows the correlation coefficient as a function of tSNR for an (a) Auditory region and (b) DMN region for each subject. The p-value for the linear regression is less than 0.05 indicating the fit statistically significant.

Fig. 4: Effect of surface mesh upsampling on representing fine-scale features. (a) The number of unique fMRI voxel for different surface upsampling factors ranging from 1 to 4, and for the original surface. Dashed lines represent these values for each individual subject; bars show the average across subjects. (b) Voxels missing from when using the original surface that were captured with increasing upsampling factors. Color indicates voxel index. (c) Thin-stripe columnar pattern detected in V2 from original surface and upsampled surface. p<0.001. Value of color bar: −log(p).

Fig. 5: Spatial nonuniformity of resolution induced by geometric distortion, in this case due to gradient nonlinearity. Voxel size varies smoothly across brain regions due to geometric expansion and compression from gradient nonlinearity, which varies with the design of the gradient coil. Examples from whole-brain axial slices (a) and histogram (b) show the spatial distribution of relative voxel size in volume dimension. Color gradient from blue to red indicate smaller to larger true voxel sizes.

Figure 3. The latent dimensions can be organized into 6 groups (shown in rows) based on their spatial similarities. Panel A shows how the spatial profile at the max-variance time (in figure 2) changes when sliding a single latent variable from -3 to +3. Panel B shows the spatial similarities among latent variables, which were clustered using K-means clustering (K = 6). The cluster number and the variance explained were also shown. Panel C shows the weighted mean functional connectivity of each cluster of latent variables.

Figure 2. Spatial temporal patterns extracted by latent dimensions. Each subplot is obtained by making one latent variable equal to +3 (+3σ for Gaussian distribution) while fixing the rest of the latent variables at zero. The x-axis is time in seconds. The y-axis is the 246 parcels. The patterns have arbitrary units, but all subplots share the same display scale so that higher variance results in higher contrast. The 32 latent variables are already organized in 6 clusters (see figure 3). The black cursor indicates the maximum variance time.

iCAPs coupling differences when the same data undergoes different temporal filtering. Red and blue denote co-activations with the same or opposite signs, respectively. A) Couplings with significantly longer duration in prep₁. B) Couplings with a significantly shorter duration in prep₁. Couplings were measured in terms of the percentage of total scanning time of the two co-occurrent iCAPs.

iCAPs total duration percentage from two pre-processing pipelines (same data). PrimVIS: primary visual, SecVIS: secondary visual, aInsula: anterior insula, Language: language network, dACC/dPFC: dorsal anterior cingulate cortex/dorsolateral prefrontal cortex, pDMN: posterior default mode network, DMN: whole default mode network AUD/SM: auditory/sensorimotor, AUD: auditory, AMY/HIP: amygdala/hippocampus, SubCort: subcortical network. Asterisk denotes significant differences (p-value_FDR< 0.05)

Figure 4. (a) Unilateral AUD seed location in healthy subject 2, (b) low-frequency resting-state connectivity map thresholded at 0.6, (c) HRAN regressed data high-frequency resting-state connectivity maps and (d) spectrally and temporally segmented regressed high-frequency resting-state connectivity map, both thresholded at 0.3.

Figure 5. Frequency power spectrum using visual seed in healthy subject 1 visualized using Turbofilt. (a) Shows the spectrum before regression, (b) shows the spectra after using HRAN and (c) is the spectrum after performing the spectrally and temporally segmented regression. The boxes correspond to labels of physiological noise. The curves correspond to tentative fits of the noise spectral power. The strong peak at 0.566Hz is a machine artifact.

Figure 1: Simulating HRFs with different temporal dynamics. A) Six HRFs with varying TTPs and FWHMs that were compared in our simulation. SPM HRF is the canonical two-gamma HRF used in SPM software while the five other HRFs have TTPs and FWHMs based on literature⁷. B) Spectrum of each HRF C) Changes in predicted fMRI response amplitude across different frequencies for different HRF timings.

Figure 3: Differences in resting state spectra across fast and slow voxels. A) Example spectrum showing how each variable shown in Fig. 3b-d were calculated. B-D) Differences in the average of the B) slopes of the resting state spectrum, C) average power, and D) exponent of fit to b-log₁₀⁡Freq^x fast and slow voxels in a session, 8/8 slope, 7/8 low frequency power, 7/8 aperiodic exponent pairings have significant difference (p<0.5), error bars report SEM.

Fig.2: Overview of the methods. (A): Preprocessed BOLD time courses from a single subject; (B): Reconstructed phase attractors in the 3D space ;(C) Feature extraction by calculating the sum of the length of the portrait edges (SE) for each attractor;(D): Calculated SE map of a subject; (E): Iteration of (A)-(D) over all subjects of the datasets; and (F): statistical analysis.

Fig.4: (a) Whole brain t-stat map for COBRE data set; (b) Whole brain t stat maps for B-snip data set.

Figure 1) DCM models. A) Model 1: included volumes of interest (VOIs) are CRBL, M1, SMA, and PMC. B) Model 2: included VOIs are FP, M1, SMA, and PMC. Between-regions connections and self-inhibitory connections for each regions are reported. Driving input representing the squeezing-ball task are set for M1 and SMA. Modulatory input representing the force level used during the task is specified for all M1 input connections.

Figure 3) Second-Level analysis outcomes. Group analysis is performed in terms of Free variational energy (F). Bayesian model selection (BMS) parameters are: posterior probability (A), protected exceedance probability (B) and Bayesian Omnibus Risk (BOR) (C). All panels confirm that model 1 is the best choice.

Figure2: example of gfc map in MNI space

Figure 1: example of GFC definition from the distribution, observed student t-score distribution (blue) and fit FWHM to a Gaussian (orange)

Figure 1. Deep linear model (shown as (c), (d)) versus shallow linear model (shown as (b)). (a) SG represents the input fMRI signal matrix, containing the t time points and m voxels. (c1) and (d1) represents the weight matrix/dictionary identified via SG of 1st and 2nd layer, respectively. (c2) and (d2) represents the feature matrix of 1st and 2nd layer, respectively. The dashed blue rectangle indicates the deeper features beyond the 2nd layer.

Figure 2. Comparison of six representative 1st layer networks from all four deep linear models (presented in the second to fifth column) with the ground truth templates (presented in the first column) from simulated fMRI data. (a) Three networks illustrate better intensity matching to the templates by Deep MF and Deep SDL than by Deep FICA or Deep NMF. (b) Three networks show better spatial matching to the templates by Deep FICA and Deep NMF than Deep MF or Deep SDL. Auditory Network: AUD. Brainstem/Cerebellum: B/C. Default Mode Network: DMN. Visual networks: VIS-1, VIS-2, VIS-3.

Figure 1. Normalized Fourier power spectra. Normalized Fourier power spectra calculated after each preprocessing step. Values have been binned and averaged across subjects. Error bars represent standard error across subjects. a) Motion Correction: Respiratory frequency, respiratory harmonics, and cardiac frequency labelled, b) Smoothing: Respiratory frequency, and respiratory harmonics labelled, c) Motion Regression: Respiratory frequency labelled, d) High Pass Filtering: no visible physiological frequency peaks.

Figure 2. Normalized Power Spectra Difference Ratios. Normalized Fourier power spectrum difference ratios expressed as a percent change. Values have been binned and averaged across subjects. Error bars represent standard error across subjects. a) Motion Correction: Respiratory frequency labelled, b) Smoothing: Respiratory frequency labelled, c) Motion Regression: Respiratory frequency labelled, d) High Pass Filtering: no visible physiological frequency peaks.

Figure 1: Overview of deep learning framework. A) The surface-based data is parameterized using spherical coordinates to a 2D “image” to serve as CNN input. B) A U-Net architecture is used to learn the relationship to the output data. Outputs are 2D “images” of anatomical or physiological features also parameterized using spherical coordinates. C) Example estimation of cortical thickness.

Figure 2: Results of the CNN predictions for an exemplar test subject. True and predicted thickness (A), MRA intensity (B) and cortical orientation angles (C). Feature maps are displayed on the inflated and pial surface. The prediction error varies spatially, but generally the pattern of error is unstructured with some focal regions exhibiting higher predictability than others. Generally, the predicted feature maps appear to be a smoothed version of the ground truth feature maps.

The BOLD responses during brains' numbering process.

A deep-learning network model for brain's number system.

Fig.1 The averaged functional connectivity matrices of MB rs-fMRI and Conventional rs-fMRI. MB rs-fMRI had a slightly stronger functional connectivity pattern in comparison to conventional acquisition sequence.

Fig.4 The distribution of brain regions, left (L) and right (R), with different nodes efficiency. The regions with significantly higher efficiency were colored in red. Reversely, in blue (P<0.05, FDR corrected). The node sizes indicate the significance of the between-group differences in regional efficiency. OLF=olfactory cortex, SFGmed=superior frontal gyrus, medial, AMYG=amygdala, MOG=Middle occipital gyrus, PreCG=precentral gyrus, SMA=supplementary motor area, SPG=superior parietal gyrus.

Figure 2. Matched seeds. Top row shows manually picked seed and bottom row shows automatically generated seed.

Figure 1. Generating feature vector by combining rs-fMRI connectivity and T1 parcellation. Panel A: Z-map, Panel B: Z-score distribution, Panel C: Freesurfer parcellation, Panel D: feature vectors.

Distribution of group parcel cohesion (mean) vs group parcel dispersion (std). Most parcels are represented as a histogram (gray scale), the largest 20 parcels are shown individually (color, size proportional to parcel size). The overall size-weighted group mean (0.58, black line) and minimum (0.5, grey line) are also shown. a) Corresponding spatial map. Distributions of cohesion shown for the current (b), anatomical (c), and connectivity-based (d) parcellations when projected to each subject’s data. GCP = group parcel cohesion, Des = Destrieux⁶, YC = Yeo/Choi^7,8.

Parcels are validated across all 18 subjects based on parcel cohesion (top row), homogeneity (middle row), and modified silhouette (bottom row) for group cohesive (left column), anatomical, and connectivity-based (right column) parcellations. Boxes represent median and interquartile ranges, whiskers cover 95% of the distributions. Group mean are also shown (black lines and bottom right values). GCP = group parcel cohesion, Des = Destrieux⁶, YC = Yeo/Choi^7,8.

Group cohesive parcellation for the group of the six subjects

DICE scores for LOOCV, random parcellation and intra-subject cross-validation

Fig. 2: Parcellation results of the primary auditory cortex for (a) left and (b) right hemisphere. Cluster 1 (black) showed the strongest SC to STG. Cluster 2 (green) was located in gyrus part of HG. Cluster 3 (red) showed strong SC to insula.

Fig. 3: T-values for regions with significantly different mean FC visualized in the atlas space. Positive t-values are presented as red color which implies the mean FC values in a pair of regions satisfy cluster1 > cluster3. Negative t-values are presented as blue color which implies the mean FC values in a pair of regions satisfy cluster1 < cluster3. The seed is located in (a) the left hemisphere and (b) the right hemisphere, respectively.

Figure 5. correlation REHO vs Respiration Rate, Heart Rate, Motion Parameters and FWHM inter-group. (a) Scatter plots of REHO (before correcting the fMRI data using BlurToFWHM) vs artifects. (b) Correlation of REHO in each voxel vs artifects across 9 datasets. (c) Scatter plots of REHO (after BlurToFWHM) vs artifects. (d) Correlation of REHO in each voxel vs artifects across 9 datasets. (e) Scatter plots of REHO (after BlurToFWHM and regression) vs artifects.

Figure 2. Correlation maps of REHO vs Respiration Rate across time inter-subject heterogeneity. (a) a scatter plot showing the correlation in the cross marked voxel. (b) Average correlations between the REHO of 10 segments (3D+time data was divided across time) and the respiratory series in each voxel of each subjects. (c) Average correlation map of 9 datasets. (d) Correlation map calculated by pooling all 4120 REHO sub-bricks together.

Figure 3. The correlation score difference between combined networks from both in Figures 1 and 2.

Figure 2. The correlation score difference between low mind-wandering and high mind-wandering networks. The score difference is significantly lower for the first block (p<2e-7 for Run1 and p<0.022 for Run 2), while they are not significant for later blocks.

Figure 2. Pearson’s correlations between a subset of the extracted QC metrics. The correlation matrix was obtained considering the whole HCP dataset.

Figure 3. The box plots show the distributions of some representative QC metrics for the two samples of LM and HM subjects (n=250 in each sample). The criteria for sample definition was based on the extreme values of the CensFDDVARS distribution (first plot) and was limited to scans with RL phase encoding.

Figure 1: Schematic flowchart of the methodology employed in this study.

Figure 3: Correlation heatmap between ICA components of both quantum microstates and resting-state fMRI networks. The correlation values are plotted in colour and mentioned numerically inside the box. Insignificant correlations are shown as zero inside the box.

Figure 1. The network formed by the MPFC and the right LP from the DMN and the left and right amygdala showing a significant interaction (Real-Sham stimulation) x (Before-After stimulation).

Figure 2. Boxplots presenting the difference (Before-After stimulation) in the connectivity between the MPFC (top) and right LP (bottom) with both amygdala for the real and sham stimulation.

Fig.1. RS-fMRI global correlations of fourteen subjects obtained at baseline (A) and at 8 weeks after the training at 15% of MIP (B) and profoundly decreased connectivity patterns at 60% of MIP

Fig. 2. Compared with Default-mode-network (DMN) activity at baseline (A), at 8 weeks after breath-training at 15% of MIP (B) reduced activation (yellow-to-red) at ACC, PCC and left intraparietal sulcus (IPS) along with diminutive activation at right IPS. Deactivation (blue) is seen in the bilateral insula and periventricular area (B). Note that reduction in BOLD activation at ACC network and at posterior cingulate-retrosplenial cortex, disappeared at the temporal lobes at 60% of MIP (C)

Figure 2. A) Coronal section of the human brain showing the representative Broadmann areas studied. B) Pairwise connectivity in ROI-to-ROI FC during rest. Strongest FC is shown in dark red. C) Mean z score (±SEM) showing FC in five strongest connections from panel (B), between BA 41 and 42, between BA 42 and 22, and between left and right hemisphere of BA 41, 42 and 22, during rest and each task condition. BA: Brodmann area. BA 41/42/22: primary/secondary/tertiary auditory cortex. NS/AS/VS: no task/acoustic task/visual task.

Figure 1. Global correlation maps represent a measure the connectedness of each voxel, characterized by the strength and sign of connectivity between a given voxel and the rest of the brain. The DMN was identified during resting-state (NS), while during task-state (AS and VS) DMN appeared to be reduced. Language network and dorsal-visual network were identified during AS and VS, respectively. Maps were thresholded at a voxel threshold of p < 0.001 (uncorrected) and a cluster threshold of p < 0.05 (FDR corrected). DMN: Default Mode Network. NS/AS/VS: no task/acoustic task/visual task.

Figure 1: Network connectome for (a) 500Hz (PTA) modulated 1-Back Memory task, (b) 1-Back Memory without modulation and (c) 1-Back recall without modulation.

Figure 2. Latent brain states identified from the HMM. Compared to the other states, S1 shows increased DMN activity while S2 shows increased DAN and SN activity. S3 appears to be a transition state consistent with that presented in literature¹¹, and S4 is prevalent during the squeeze blocks of the paradigm. DMN - default mode network; FPCN - fronto-parietal control network; DAN - dorsal attention network; SN - salience network; LC - locus coeruleus.

Figure 3. Boxplot showing the average pupil dilation during switching between brain states for active session (left, blue) and sham session (right, red).

Table 1. Network interaction across different brain states in alphabet Navon task

Figure 1. Intermediate frequency modulation for recall condition in the frequency range (a) 10 to 16 kHz and (b) 1 to 3 kHz.

Figure 2. Frequency modulation for (a) memory condition at 2 kHz (b) recall condition in the intermediate range of 1 to 3 kHz.

Figure1 KIBRA rs17070145 interacts with gender on brain gray matter volume and long-range functional connectivity density. Male KIBRA C-allele carriers showed greater GMV in the inferior temporal gyrus, while decreased lrFCD in the left middle temporal gyrus and left middle cingulate gyrus than male TT homozygote. FWE correction, voxel p < 0.01, cluster p < 0.05.

Figure 2. (A) Cerebellum mask (B &C) DC changes in cerebellum (8,9) is correlated with sleep improvement (D) Insignificant DC changes between verum and sham groups in cerebellum (8,9)

Figure 3. (A) Functional connectivity (FC) between the cerebellum (8,9) seed and the thalamus (B & C) FC changes in thalamus are correlated with sleep improvement in both groups (D) Significant FC changes between the verum and sham groups in the thalamus (**p < 0.01).

Figure 2. Left column indicates the FNC matrices of States 1-4, with the number of windowed FNCs in every state, the corresponding percentage, as well as the number of participants from every group who entered into the state. Right column represents the visualization of functional network connectivity in each state. The functionalconnectivity matrix was screened using a threshold of 0.1 to display all independent components of functional networks. HC, healthy control; PA, cirrhotic patient.

Figure 3. Significant group differences in the functional network connectivity in each state. The results are displayed as the -sign(t)*log10(p). Statistical threshold wasestablished at an FDR-corrected P<0.05. HC, healthy control; PA, cirrhotic patient.

Fig. 1: Differences of the cortical-striatal FC in patients with HFS compared to healthy controls. The FC between striatal subregions and both motor and emotion-related cortex was statistically different between the two groups (GRF correction, voxel P < 0.005, cluster P < 0.05). The yellow dots in the brain represent the seeds of striatal subregions. The red regions represent the increased FC between seeds and cortexes, while the blue areas represent a decreased FC. The color bar represents the t value.

Figure 1. Within-group ALFF maps in the healthy control (HC) group as well as the patients infected with hepatitis C virus (HCV). The letters “L” and “R” represent the left and right sides, respectively.

Figure 3. The functional connectivity pattern of seed region (i.e. left medial frontal gyrus and bilateral anterior cingulate gyrus) in the healthy control (HC group) and the patients with Hepatitis C Virus infection (HCV group). Compared with HC group, HCV group shows decreased functional connectivity between seed region and right middle frontal gyrus. The letters “L” and “R” represent the left and right sides, respectively.

Figure4 Intergroup differences in internetwork FC between the tinnitus patients and HCs and between the patients before and after sound therapy.

(A) Compared with the HCs, the tinnitus patients exhibited a decreased (i.e., positive) (mVN and lVN) or an increased (i.e., less negative) (mVN and SMN; mVN and AN) internetwork FC at baseline.

Figure4 (B) Compared with HCs, tinnitus patients showed a decreased (i.e., positive) (AN and pDMN; DAN and LFPN) internetwork FC after sound therapy.

The brain networks of the LBP group lead to the slower information processing in the brains of those with LBP compared to PHN.

Fig. 1 Statistical nonparametric map showing regions with significant decrease in functional connectivity (FC) with the ACC after 10days of CLAV administration compared to baseline. Regions of decreased FC (FDR=0.01, corrected) in blue-green scale (paired t-test T-value threshold 5.3) overlaid on the standard MNI-152 brain in grey scale. Right side sagittal slice (y=-56) shows the angular gyrus region. Left side axial slice (z=33) shows the angular gyrus, precuneus, and posterior cingulate cortex regions. R=right L=left A=anterior P=posterior I=inferior S=superior

Fig. 2 Statistical nonparametric map showing regions with significant increase in functional connectivity (FC) with the ACC after 10days of CLAV administration compared to baseline. Regions of increased FC (FDR=0.01, corrected) in red-yellow scale (paired t-test T-value threshold 5.3) overlaid on the standard MNI-152 brain in grey scale. Right side sagittal slice (y=-6) shows the Paracentral gyrus (pre and post-central) regions. Left side axial slice (z=48) shows the Supplementary Motor Area and dorsal ACC regions. R=right L=left A=anterior P=posterior I=inferior S=superior

Figure 2. Aβosa and Aβice induce seed-based analysis of connectivity changes amongst network. Whole-brain dictionary learning maps depicting 19 cortical and subcortical components found in 4 main networks (upper). Group differences seed-based connectivity pattern through components for inoculum site (dentate gyrus)(lower). Asterisk represents significant p-value (p<0.001) of the group difference in Fisher z-transformed correlation.

Figure 5. Decreased glutamate levels after Aβosa inoculation in APPswe/PS1de9 mice. a. GluCEST average signal evolution in APPice, APPosa and APPcontrol mice shows age-related modification detected in 9 mpi animals. They are rescued by Aβice inoculation. b. Variation maps of GluCEST at 4mpi (right). Variations was calculated in every region from the atlas. Colors represent hyposignal (blue) and hypersignal (red) comparing to control mice (*p<0.05).

Figure 1. (i) Regions wherein GMV are significantly increased after acupuncture. The centers of 3 ROIs are MNI space coordinates of these clusters’ peak voxel (ROI 1: (-38, -72, -35); ROI 2: (-5, -54, 8); ROI 3: (39, -41, 66)). (ii) Regions wherein GMV are significantly increased after diet control therapy. The centers of 2 ROIs are the MNI space coordinates of these clusters’ peak voxel (ROI 4: (18, -9, -38); ROI 5: (-30, 47, 27)). P: Posterior; A: Anterior. Results (scan 2 – scan 1) are reported with a threshold of p < 0.001 (uncorrected) on the voxel level and a cluster extent > 30 voxels.

Figure 2. The results of FC with 5 ROIs after acupuncture. A: Row 1 is the region wherein FC with ROI 2 are significantly increased and row 2 is the region wherein FC are decreased. B: The region wherein FC with ROI 3 are significantly decreased. C: The region wherein FC with ROI 4 are significantly increased. D: The region wherein FC with ROI 5 are significantly increased. L: Left; R: Right. Results (scan 2 – scan 1) are reported with a threshold of p < 0.001 (uncorrected) on the voxel level and a cluster extent ≥ 5 voxels.

Figure 1. Different brain regions compared by ALFF value of short-term and long-term abstinence groups