Predicting brain age of children accurately and quantitatively can give help in brain development analysis and brain disease diagnosis. Traditional methods to estimate brain age based on 3D magnetic resonance (MR) T1 weighted imaging (T1WI) and diffusion tensor imaging (DTI) need complex preprocessing and extra scanning time, decreasing clinical practice, especially in children. This research aims at proposing an end-to-end AI system based on deep learning to predict the brain age based on routine brain MR imaging. We spent over 5 years enrolling 220 stacked 2D routine clinical brain MR T1-weighted images of healthy children aged 0 to 5 years old and randomly divided those images into training data including 176 subjects and test data including 44 subjects. Data augmentation technology, which includes scaling, image rotation, translation, and gamma correction, was employed to extend the training data. A 10-layer 3D convolutional neural network (CNN) was designed for predicting the brain age of children and it achieved reliable and accurate results on test data with a mean absolute deviation (MAE) of 67.6 days, a root mean squared error (RMSE) of 96.1 days, a mean relative error (MRE) of 8.2%, a correlation coefficient (R) of 0.985 and a coefficient of determination (R2) of 0.971. Specially, the performance on predicting the age of children under 2 years old with a MAE of 28.9 days, a RMSE of 37.0 days, a MRE of 7.8%, a R of 0.983 and a R2 of 0.967 is much better than that over 2 with a MAE of 110.0 days, a RMSE of 133.5 days, a MRE of 8.2%, a R of 0.883 and a R2 of 0.780.
Objective:Rabbit VX2 hepatoma was treated with transcatheter hepatic artery infusion with recombinant human(rh)endostatin combined with transcatheter arterial chemoembolization(TACE).The therapeutic efficacy was evaluated and the peri-cancerous CT perfusion features as well as the expressions of microvascular density(MVD) and vascular endothelial growth factor(VEGF) were analyzed.Methods:Thirty rabbits with VX2 hepatoma were randomly divided into three groups with 10 rabbits in each group.Group A were treated with TACE and infused with rh-endostatin via hepatic artery.Group B were treated with TACE only.Group C were control group.Routine CT scanning was performed before therapy and CT perfusion imaging were performed at 2nd weeks after treatment to get the data of blood flow(BF),blood volume(BV) and capillary permeability surface(PS) area.The rabbits were sacrified immediately after scanning.The MVD and VEGF expressions in tumor tissue slides were detected by immunohistochemistry.Results:The hepatoma growth rate was significantly lower in group A and B than that in group C(P0.01).The BF,BV,and PS were significantly increased and MVD and VEGF were significantly decreased in group A than those in group B and C(P0.05,P0.01).There was no significant difference in the value of BF,BV,PS,MVD,and VEGF between groups B and C(P0.05).The expression of MVD positively correlated with the expression of VEGF in linear dependency(P0.01).BF,BV,and PS had positive correlation with MVD and VEGF in group A(P0.05).BF,BV,and PS had no significant correlation with MVD and VEGF expression in group B(P0.05).Conclusion:The combination therapy of TACE and administration of recombinant human endostatin via hepatic artery infusion significantly slows down the growth speed of hepatoma and inhibits tumor angiogenesis.The parameters of CT perfusion imaging can not reflect the changes of pathological parameters such as MVD and VEGF at any therapeutic condition.
Objective:To investigate the correlation between scanning time and collimation in lower limb MSCT angiography.Methods:70 patients suspected of having lower limb arterial occlusion were divided into the experimental group(n=35) and the control group(n=35),both receiving CTA examination.16×1.5mm collimation was used for the experimental group and 16×0.75mm collimation for the control group.The scan ranges were both 1100~1200mm and the scan paramters of the two groups were the same.The scanning time,volume of contrast material and the coverage on z-axis were analyzed statistically.Result:Using two different collimations,the scanning time in the experimental group and the control group was(18.13±5.12)s and(31.43±3.12)s respectively,the volume of contrast material was(80±10)ml and(120±20)ml respectively,and the coverage on z-axis was 24mm in the experimental group,doubling that in the control group.Conclusion:The 16×1.5mm collimation can markedly reduce the scanning time and volume of contrasts material,in the mean time it can extend the coverage on z-axis.
Objective To investigate how functional connectivity changes within default-mode network related to posterior cingulated cortex employing resting-state functional MRI (tMRI). Methods fMRI was compared between 16 mild Alzheimer' s disease (AD) patients and 16 normal elder subjects. Regions of functional connectivity to posterior cingulated cortex were gathered by calculating temporal correlations in low frequency fMRI signal fluctuations. SPM2 was applied to calculate significant differences of connectivity between group and within group. Significance threshold was set up at the corrected P 5. A random effect two-example t test was performed by SPM2 to achieve significant difference of functional connection between groups ( P 5 ). Results Regions showing disrupted connectivity to posterior cingulated cortex were: ventral medial prefrontal cortex (MPFC), bilateral visual cortex, infero-temporal cortex (ITC), and left hippocampus, right thalamus, right dorsal-lateral prefrontal cortex ( DLPFC), and precuneus. There were also regions showing increased connectivity with leftward asymmetry, these regions included: MPFC, left ITC, bilateral DLPFC, left pre- central motor cortex and left basal ganglia. Conclusions Impairments of memory and high visual-related functions in AD can be explained by functional disconnection in resting-state. Remoldability is reserved in mild AD to compensate for brain function which is taxed by left hemisphere preferentially. Our findings suggest that resting-state fMRI might be an appropriate approach for evaluating AD brain mechanism.
Key words:
Alzheimer disease; Magnetic resonance imaging; Gyrus cinyuli
(Aim) Sensorineural hearing loss (SNHL) is correlated to many neurodegenerative disease. Now more and more computer vision based methods are using to detect it in an automatic way. (Materials) We have in total 49 subjects, scanned by 3.0T MRI (Siemens Medical Solutions, Erlangen, Germany). The subjects contain 14 patients with right-sided hearing loss (RHL), 15 patients with left-sided hearing loss (LHL), and 20 healthy controls (HC). (Method) We treat this as a three-class classification problem: RHL, LHL, and HC. Wavelet entropy (WE) was selected from the magnetic resonance images of each subjects, and then submitted to a directed acyclic graph support vector machine (DAG-SVM). (Results) The 10 repetition results of 10-fold cross validation shows 3-level decomposition will yield an overall accuracy of 95.10% for this three-class classification problem, higher than feedforward neural network, decision tree, and naive Bayesian classifier. (Conclusions) This computer-aided diagnosis system is promising. We hope this study can attract more computer vision method for detecting hearing loss.
Full text Figures and data Side by side Abstract eLife digest Introduction Results Discussion Materials and methods References Decision letter Author response Article and author information Metrics Abstract Hearing loss often triggers an inescapable buzz (tinnitus) and causes everyday sounds to become intolerably loud (hyperacusis), but exactly where and how this occurs in the brain is unknown. To identify the neural substrate for these debilitating disorders, we induced both tinnitus and hyperacusis with an ototoxic drug (salicylate) and used behavioral, electrophysiological, and functional magnetic resonance imaging (fMRI) techniques to identify the tinnitus–hyperacusis network. Salicylate depressed the neural output of the cochlea, but vigorously amplified sound-evoked neural responses in the amygdala, medial geniculate, and auditory cortex. Resting-state fMRI revealed hyperactivity in an auditory network composed of inferior colliculus, medial geniculate, and auditory cortex with side branches to cerebellum, amygdala, and reticular formation. Functional connectivity revealed enhanced coupling within the auditory network and segments of the auditory network and cerebellum, reticular formation, amygdala, and hippocampus. A testable model accounting for distress, arousal, and gating of tinnitus and hyperacusis is proposed. https://doi.org/10.7554/eLife.06576.001 eLife digest One in three adults over the age of 65 will experience a significant loss of hearing. This is often worsened by related conditions, such as: tinnitus, an unexplained constant buzzing or ringing sound; and hyperacusis, whereby everyday sounds are perceived as too loud or painful. Most hearing loss is caused by damage to the sound-sensitive cells within a structure in the inner ear called the cochlea. Some studies have also identified regions of the brain that show abnormal activity in people with tinnitus and hyperacusis. However, the results from different patients have often been inconsistent and sometimes contradictory, and so it remains unclear what exactly causes these conditions. To overcome this problem, Chen et al. made use of the fact that tinnitus and hyperacusis are common short-term side effects of certain drugs and measured the brain activity in rats before and after they were given one such drug. Before receiving the drug, the rats had first been trained to expect to receive a food pellet from the left side of their cage when they heard a steady buzzing sound. The rats were also trained to expect a food pellet from their right if they heard nothing at all. Shortly after receiving the drug, the rats often failed to respond correctly in the ‘quiet tests’ and behaved like they were already experiencing a constant buzzing sound, as would be expected if they had tinnitus. Further tests confirmed that the drug also triggered behavior in the rats that is typical of people with hyperacusis. Chen et al. then discovered that the drug treatment reduced the nerve signals that are sent from a rat's cochlea. Moreover, the drug treatment greatly increased the activity in response to sound within parts of the rat's brain; these and other parts of the brain also became overactive in drug-treated rats in the absence of sound. Finally, further experiments revealed that drug-treated rats had stronger connections between these brain regions than in normal rats. Chen et al. used these results to propose a model to explain the underlying causes of tinnitus and hyperacusis. However, because the drug treatment only induces short-term hearing impairment, further studies are needed to see if this model also applies when these conditions are long-term. https://doi.org/10.7554/eLife.06576.002 Introduction A third of adults over the age of 65 suffer from significant hearing loss, a condition exacerbated by two debilitating condition, subjective tinnitus, a phantom ringing or buzzing sensation, and hyperacusis, normal sounds perceived as intolerably loud or even painful. Roughly 12% of adults experience tinnitus, but the prevalence skyrockets to 50% in young combat personnel (Leske, 1981; Andersson et al., 2002; Cave et al., 2007; Michikawa et al., 2010; Hebert et al., 2013). Tinnitus is costly with more than $2 billion paid annually in veteran disability payments. Hyperacusis affects roughly 9% of adults (Andersson et al., 2002), but its prevalence is likely higher because of the difficulty of self-diagnosis (Gu et al., 2010). Remarkably, among those whose primary complaint is hyperacusis, 90% also suffer from tinnitus (Baguley, 2003). Since tinnitus and hyperacusis are often triggered by cochlear hearing loss, it was long assumed that these auditory distortions resulted from hyperactivity disorders in the peripheral auditory nerve. This hypothesis, however, is contradicted by studies showing that auditory nerve spontaneous and sound-evoked firing rates are depressed in subjects with cochlear damage (Kiang et al., 1970; Wang et al., 1997). Moreover, surgical section of the auditory nerve fails to eliminate tinnitus (Baguley et al., 1992; Lockwood et al., 2001). These negative results plus recent imaging studies now suggest that tinnitus and hyperacusis arise from maladaptive neuroplastic change in the central nervous system (CNS) provoked by cochlear pathology (Lockwood et al., 1998; Husain et al., 2011; Sereda et al., 2011). Several models of tinnitus and hyperacusis have been proposed that involve increased central gain, altered functional connectivity (FC), and aberrant neural oscillations in neural networks (Weisz et al., 2007; Sereda et al., 2011; Henry et al., 2014). Most of these conceptual models have emerged from human imaging studies using magnetoencephalography, electroencephalography, magnetic resonance imaging (MRI), and functional MRI (fMRI) of the blood oxygen level-dependent (BOLD) response (Llinas et al., 1999; Weisz et al., 2005; Auer, 2008; Gu et al., 2010; Moazami-Goudarzi et al., 2010; Leaver et al., 2012; Maudoux et al., 2012; Husain and Schmidt, 2014). In the context of central gain models, some human imaging data indicate that hyperacusis is associated with enhanced sound-evoked activity in multiple-auditory processing centers, namely auditory cortex (ACx), medial geniculate body (MGB), and inferior colliculus (IC), whereas tinnitus can be triggered solely by enhanced central gain in the ACx (Gu et al., 2010). On the other hand, active loudness models suggest that tinnitus arises entirely from increased central noise independent of gain, whereas hyperacusis results exclusively from increased nonlinear gain that results in loudness intolerance (Zeng, 2013). While cross-sectional human brain imaging studies have identified many different sites of aberrant neural activity, published results from patients have often produced diverse, inconsistent, or contradictory findings. Some discrepancies are likely due to confounding factors such as patient heterogeneity, unknown etiology, genetic diversity, social and environmental factors, and duration or severity of tinnitus and hyperacusis. Animal models could potentially overcome many of these limitations provided that tinnitus and hyperacusis can be reliably induced, behaviorally measured, and functionally imaged. While tinnitus can develop in some individuals after intense noise exposure, the percentage of affected individuals is highly variable and its duration is unpredictable (Heffner and Harrington, 2002; Lobarinas et al., 2006; Heffner, 2011). High doses of aspirin, an anti-inflammatory drug used to treat rheumatoid arthritis, have long been known to consistently induce acute tinnitus in humans and animals (Myers and Bernstein, 1965; Myers et al., 1965; Mongan et al., 1973). Moreover, high-dose sodium salicylate (SS), the active ingredient in aspirin, not only consistently induces tinnitus (Jastreboff et al., 1988; Lobarinas et al., 2004; Stolzberg et al., 2013), but also hyperacusis (Chen et al., 2014; Hayes et al., 2014); these perceptual disorders disappear a day or two post-treatment. The highly predictable time course of SS-induced tinnitus and hyperacusis makes it an extremely powerful tool for studying the neural correlates of these perceptual disturbances. Therefore, we took advantage of our unique behavioral techniques for assessing SS-induced tinnitus and hyperacusis in rats and combined this with focused electrophysiological measurements plus global fMRI assessment techniques to map out the regions of neural hyperactivity and enhanced FC that characterize the tinnitus–hyperacusis network. To identify regions of heightened or depressed spontaneous neural activity, we measured the amplitude of low-frequency fluctuations (ALFF) in resting-state fMRI (Zang et al., 2007; Zhang et al., 2010; Yao et al., 2012; Wen et al., 2013) and combined this with resting-state FC to identify regions of increased or decreased functional coupling between regions of the auditory pathway and other parts of the CNS. This is the first animal study to use ALFF and FC combined with detailed electrophysiological measures to provide a comprehensive neurological map of the tinnitus–hyperacusis network. Results Three complimentary experiments involving behavioral, electrophysiological, and functional imaging were conducted in separate groups of rats. In Experiment 1, three behavioral studies were performed on separate groups of rats to assess SS-induced tinnitus, hyperacusis, and startle reflex hyperactivity. In Experiment 2, electrophysiological measurements were carried out on a separate group of rats to determine how SS altered the neural input/output functions in cochlea, as reflected in the compound action potential (CAP) from the auditory nerve, and the local field potentials (LFP) recorded in the MGB, ACx, and lateral amygdala (LA). In Experiment 3, resting-state fMRI studies were conducted in another group of rats to determine how SS altered the ALFF and FC patterns obtained with seeds placed in ACx, MGB, and IC. Experiment 1 Tinnitus To determine if SS-induced tinnitus, we tested three rats using our 2AFC-tinnitus paradigm. All three rats developed tinnitus-like behavior; data from two representative animals are shown in Figure 1A,B. During baseline testing (B1–B4; B6–B9), rats correctly identified Quiet (no sound stimulus) trials at greater than 70% correct and AM and NBN trials >80% correct. The saline-control treatment had no noticeable effect on performance during Quiet, NBN, or AM trials. However, when the rats were treated with SS, performance dropped to 50% or less only on Quiet trials, that is, rats shifted their response preference from the feeder associated with Quiet to the feeder associated with a continuous NBN, behavior indicative of tinnitus. When SS treatment was discontinued (P1–P4), performance on Quiet trials returned to baseline indicating that tinnitus had disappeared. Performance on NBN and AM trials was unaffected indicating that behavior was under sound stimulus control. Figure 1 Download asset Open asset SS-induced tinnitus, hyperacusis, and startle reflex hyperactivity. (A and B) 2AFC-tinnitus task for two representative rats during baseline days B1–B4, during Saline (SAL) treatment, during baseline days B6–B9, 2 hr post-sodium salicylate (SS), and days P1–P4 post-SS treatment. Percent correct performance shown for NBN, AM, and Quiet trials. Purple-shaded region is the 99% confidence interval for baseline measurements (B1–B4; B6–B9). (C) Mean (+SEM, n = 7) reaction time-intensity functions measured at baseline, after Saline-treatment and after SS-treatment. Reaction times during SS treatment were significantly longer than baseline at 30 dB SPL and significantly shorter than baseline at 50–90 dB SPL (*p < 0.05; **p < 0.01; and ***p < 0.001). (D) Mean (+SEM, n = 6) startle amplitude-intensify functions after treatment with Saline or SS. Startle amplitudes after SS treatment were significantly larger than after Saline at 95 and 105 dB SPL (p < 0.001). https://doi.org/10.7554/eLife.06576.003 Hyperacusis To test for SS-induced hyperacusis, we measured reaction time-intensity functions to broadband noise bursts before and after Saline or SS-treatment (Figure 1C). Reaction time–time intensity functions obtained with Saline were nearly identical to those obtained during baseline indicating that the injection had no effect on behavior. In contrast, reaction times obtained after SS were significantly different from baseline at low and high intensities [Two-way, repeated measures ANOVA; significant effect of treatment (p < 0.0001), intensity(p < 0.0001), interaction of treatment × sound intensity (p < 0.001); Bonferroni post-hoc analysis between baseline and SS significant at 30 dB (p < 0.001), 50 dB (p < 0.05), 60 dB (p < 0.01), 70 and 80 dB (p < 0.001), and 90 dB (p < 0.05)]. Reaction times after SS were significantly shorter than baseline at moderate to high intensities behavioral evidence indicative of hyperacusis, that is, these intensities were perceived as louder than normal (Lauer and Dooling, 2007; Chen et al., 2014; Hayes et al., 2014). However, at 30 dB SPL (sound pressure level) reaction times were longer than normal due to hearing loss, which reduces the loudness of sounds near threshold. Reaction time–intensity functions returned to normal after SS treatment was discontinued (data not shown). Startle reflex hyperactivity Startle reflex hyperactivity has been linked to hyperacusis (Sun et al., 2009; Lu et al., 2011). Therefore, acoustic reflex amplitude-intensity functions were compared in the same group of rats after Saline and SS treatments (Figure 1D). SS treatment caused a significant increase in startle amplitude at 95 and 105 dB SPL [Two-way, repeated measure, ANOVA, significant main effect of intensity (p < 0.001), treatment (p < 0.001), intensity × treatment interaction (p < 0.001); Bonferroni post-test significant at 95 dB and 105 dB (p < 0.001)]. Startle amplitudes returned to normal when SS was discontinued (data not shown). Experiment 2 Electrophysiology SS is known to cause temporary hearing loss and reduce the neural output of the cochlea. To quantify the effects, CAP amplitude-intensity functions were measured before and 2 hr post-SS. The mean (+SEM) CAP amplitude-intensity function (average of 6, 8, 12, 16, 20, 24, 30, and 40 kHz) measured 2 hr after SS treatment was shifted to the right at low intensities due to a threshold shift of approximately 20 dB (horizontal arrow, Figure 2A). In addition, the amplitude of the CAP was greatly reduced (70–80%) at suprathreshold intensities (down arrow, Figure 2A) indicating a profound reduction in the neural output of the cochlea. LFP amplitude-intensity functions were also recorded from the MGB, ACx, and LA before and 2 hr after SS treatment. The LFP amplitude-intensity functions from all three structures were shifted to the right approximately 20 dB (Figure 2B–D) 2 hr post-SS consistent with the CAP. These results indicate that the threshold shift measured in central structures is largely determined by the loss of sensitivity at the cochlea. LFP amplitudes in the MGB, ACx, and LA increased rapidly with increasing intensity, and response amplitudes became substantially larger than pre-treatment values (Figure 2B–D) in contrast to the large CAP amplitude reductions (∼70% decrease) (Figure 2A). The SS-induced enhancement of suprathreshold LFP amplitudes at 100 dB SPL was approximately 50% in the MGB and 140% in the ACx, results indicative of a progressive increase in gain from peripheral to more central auditory loci (Noreña, 2010; Lu et al., 2011). Figure 2 Download asset Open asset SS depresses cochlear potentials but enhances central auditory evoked responses. Effects of 300 mg/kg SS on peripheral and central electrophysiological measures. (A) Mean (+SEM, n = 5) compound action potential (CAP) input/output function (average of 6, 8, 12, 16, 20, 24, 30, and 40 kHz; 10-ms tone burst) recorded from the round window pre- and 2 hr post-SS. Note, 20 dB threshold shift of the function to the right at low intensities (horizontal arrow) and large reduction in CAP amplitude at high intensities (down arrow). (B, C, D) Local field potential input/output functions (50-ms noise bursts) from medial geniculate body, auditory cortex, and lateral amygdala (AMY), respectively, before and 2 hr post-SS. Note, 20 dB threshold shift of the functions to the right at low intensities (horizontal arrows) and large increase in response amplitude (up arrow) at suprathreshold levels (>60 dB SPL). https://doi.org/10.7554/eLife.06576.004 Experiment 3 ALFF To identify the global effects of SS on brain activity, we compared the ALFF in the SS group with the Saline group 2 hr post-treatment using two-sample t-tests corrected for multiple comparisons. Figure 3 shows the regions where significant increases or decreases in ALFF were observed due to SS; Table 1 shows the cluster sizes and t-values in left and right hemispheres for each region. Within the cerebellum, SS produced significant bilateral increases in ALFF in the parafloccular lobes (PFL, 38–37 voxels) and cerebellar lobules 4 (CB4, 38–37 voxels) (Figure 3A,B). Significant bilateral increases in ALFF were also observed in subcortical areas that included the gigantocellular reticular nucleus/oral region of the pontine reticular nucleus (RN/PnO, 35–28 voxels, Figure 3C) and pontine reticular nucleus/paramedian raphe nucleus (PnO/PMr, 12–10 voxels, Figure 3D). In the midbrain, significant bilateral increases in ALFF occurred in the IC, a binaural auditory processing area (IC, 72–68 voxels, Figure 3C,D) (Skaliora et al., 2004) and superior colliculus (SC, 40–43 voxels, Figure 3D), a visual center with multisensory inputs (Wallace et al., 1993). Significant bilateral increases also occurred in the MGB (MGB, 52–48 voxels), a high-level auditory processing area (Figure 3E–G), in the ACx (ACx, 167–178 voxels, Figure 3E–H), visual cortex (VCx, 32–39 voxels, Figure 3C,D), somatosensory cortex (SSCx, 37–40 voxels, Figure 3I–K), and amygdala (AMY, 61–60 voxels, Figure 3E–H). In contrast, SS produced significant bilateral decreases in ALFF within the hippocampus (HIP, 108–92 voxels, Figure 3E–I) and caudate–putamen (CPu; 72–71 voxels, Figure 3I–K). Figure 3 Download asset Open asset SS enhances and depresses amplitude of low-frequency fluctuations (ALFF) in specific CNS regions. Panels A (most caudal) through K (most rostral) show MR images of rat brain. Significant differences in ALFF between the SS group vs Saline group 2 hr post-treatment. Thresholds set at a corrected p value of <0.001 determined by Monte Carlo simulation. CB4, lobules 4 of cerebellum; PFL, parafloccular lobe of cerebellum; RN, gigantocellular reticular nucleus; PnO, pontine reticular nucleus oral; PMr, paramedian raphe nucleus; VCx, visual cortex; IC, inferior colliculius; SC, superior colliculius; MGB, medial geniculate body; ACx, auditory cortex; HIP, hippocampus; AMY, amygdala; SSCx, somatosensory cortex; Cpu, caudate putamen. Color heat map scale in lower right shows corrected t-values ranging from +4.06 to −4.02. https://doi.org/10.7554/eLife.06576.005 Table 1 SS-induced changes in amplitude of low-frequency fluctuations (ALFF); SS group vs Saline group; p < 0.001 corrected for multiple comparisons https://doi.org/10.7554/eLife.06576.006 Brain regionLeftRightCluster sizet-valueCluster sizet-valueALFF increased ACx1673.8121783.746 IC723.383683.473 MGB523.432483.339 SSCx373.383403.402 VCx323.342393.312 SC404.123434.094 AMY614.192603.923 RN/PnO353.249283.290 PnO/PMr123.498103.313 PFL383.349373.292 CB4383.349373.292ALFF decreased HIP108−4.08792−4.002 CPu72−3.77271−3.741 Abbreviations: auditory cortex (ACx), inferior colliculus (IC), medial geniculate body (MGB), somatosensory cortex (SSCx), visual cortex (VCx), superior colliculi (SC), amygdala (AMY), gigantocellular reticular nucleus (RN), pontine reticular nucleus oral (PnO), paramedian raphe nucleus (PMr), parafloccular lobe of cerebellum (PFL), cerebellum lobule 4 (CB4), hippocampus (HIP), caudate-putamen (CPu), sodium salicylate (SS). Functional connectivity To determine if SS altered FC, two-sample t-tests were computed to identify regions where significant differences occurred in the FC maps for SS and Saline conditions. As shown in Figure 4 and Table 2 (cluster size, t-values shown for left and right hemispheres), when the seed region was in the ACx, there were significant bilateral increases of FC in large clusters located in the MGB (62–70 voxels), IC (82–88 voxels), and AMY (67–68 voxels) plus moderate clusters located in the RN (52–48 voxels), PFL (39–35 voxels), and CB4 (51–53 voxels). No decreases in FC were observed. When the seed regions were located in MGB, there were significant bilateral increases in FC in large clusters in the ACx (195–210 voxels) and the HIP (162–178 voxels). Again, no decreases in FC were seen. Finally, with the seeds located in the IC, there were significant bilateral increases in FC in large clusters located in the MGB (72–60 voxels) and HIP (140–131 voxels); again no decreases in FC occurred. The Figure 4—figure supplement 1 showed the BOLD data (ALFF and FC) for baseline and after SS application separately. Figure 4 with 1 supplement see all Download asset Open asset SS alters functional connectivity (FC) in specific brain resions. ROI FC heat maps showing the regions of the brain where SS induced a statistically significant increase in FC with the ROI placed in the ACx (top row), MGB (middle row), or inferior colliculus (IC) (bottom row). Scale bar shown in lower right; corrected t-values ranged from +4.06 to −2.00. CB4, lobules 4 of cerebellum; PFL, parafloccular lobe of cerebellum; RN, gigantocellular reticular nucleus; IC, inferior colliculius; MGB, medial geniculate body; ACx, auditory cortex; HIP, hippocampus; AMY, amygdala. https://doi.org/10.7554/eLife.06576.007 Table 2 SS-induced increases in functional connectivity (FC); p < 0.001 corrected for multiple comparisons https://doi.org/10.7554/eLife.06576.009 Seed regionBrain regionLeftRightCluster sizet-valueCluster sizet-valueACxMGB623.900703.859IC823.912883.632AMY673.897683.839RN523.983483.902PFL393.992353.954CB4514.066534.070MGBACx1954.0742104.084HIP1624.0321784.109ICMGB724.098604.013HIP1404.0641313.904 Abbreviations: medial geniculate body (MGB), inferior colliculus (IC), amygdala (AMY), reticular nucleus (RN), parafloccular lobe of cerebellum (PFL), cerebellum lobule 4 (CB4), auditory cortex (ACx), hippocampus (HIP), sodium salicylate (SS). Discussion Brain gain SS induced a peripheral threshold shift of approximately 20 dB for the CAP (Figure 2A) (Chen et al., 2013, 2014). The same amount of threshold shift occurred at higher levels of the auditory pathway indicating that the SS-induced hearing loss originates in the cochlea and is relayed centrally. SS also reduced the CAP neural output by ∼70% at suprathreshold intensities. Paradoxically, suprathreshold LFP amplitudes in the MGB, ACx, and LA were larger than normal despite the massive reduction in the output of the cochlea (Figure 2B–D). These provocative findings provide compelling evidence for an increase in central gain, a form of homeostatic plasticity implicated in tinnitus and hyperacusis (Salvi et al., 1990; Auerbach et al., 2014). The enhanced LFPs seen in ACx are consistent with the enhanced fMRI response observed in the ACx of tinnitus patients, whereas the enhanced LFPs seen in both ACx and MGB are consistent with the enhanced fMRI responses observed these regions in hyperacusis patients (Gu et al., 2010). These results are consistent with previous models and data linking tinnitus and loudness intolerance to increased central gain in the central auditory pathway in particular regions from the IC to ACx (Salvi et al., 1990; Qiu et al., 2000; Auerbach et al., 2014). In some models, enhanced central gain amplifies central neural noise resulting in tinnitus (Noreña, 2010). However, in other models, central neural noise increases independent of central gain (Zeng, 2013); this could potentially explain why some patients only experience tinnitus, but not hyperacusis (Baguley, 2003). However, this distinction is clouded by the fact that many tinnitus patients are unaware of their mild hyperacusis, that is, hyperacusis may be more prevalent in tinnitus patients than currently believed because many patients are unaware of their hyperacusis (Gu et al., 2010). Many cellular mechanisms could enhance central gain, but one likely candidate is reduced inhibition (disinhibition). Considerable evidence exists for dysregulated inhibition in central gain-control models (Auerbach et al., 2014). First, SS can suppress GABA-mediated inhibition and enhance excitability (Xu et al., 2005; Gong et al., 2008). Second, SS enhances sound evoked activity in ACx when given systemically or applied locally to the LA or ACx, whereas it depresses ACx responses when only applied to the cochlea (Sun et al., 2009; Chen et al., 2012). Third, drugs that enhance GABA-mediated inhibition, prevent SS-induced gain enhancement (Sun et al., 2009; Lu et al., 2011), and suppress tinnitus (Brozoski et al., 2007b). Behaviorally, hyperacusis was initially observed at 50 dB SPL; the same low intensity at which sound-evoked neural hyperactivity occurred in the ACx. In contrast, sound-evoked hyperactivity occurred at noticeably higher intensities for the LA (∼60 dB SPL), MGB (∼70 dB SPL), and acoustic-startle reflex amplitude (∼95 dB SPL). These results suggest that neural responses from the ACx may be one of the most sensitive biomarkers of hyperacusis (Juckel et al., 2004; Gu et al., 2010). However, since neural responses increased in magnitude from cochlea to cortex, loudness intolerance issues likely result from multiple stages of neural amplification as signals are relayed rostrally from the cochlea to the ACx (Auerbach et al., 2014). Indeed, there is growing evidence that the neural amplification gradually develops in the auditory brainstem and serially accumulates to supernormal levels after reaching the MGB and ACx consistent with previous electrophysiological results (Qiu et al., 2000; Schaette and McAlpine, 2011). Tinnitus Some models of tinnitus are based on changes in spontaneous spiking patterns such as increased firing rate or increased neural synchrony (Eggermont, 2015). SS either decreased or had no effect on spontaneous spike rate in primary ACx (Ochi and Eggermont, 1996; Yang et al., 2007) and reportedly no effect on synchrony between neuron pairs (Eggermont, 2015). Since the BOLD and LFP responses mainly represent presynaptic activity, it is difficult to directly relate our results to these spiking models. However, the increase in very low-frequency BOLD oscillations (0.01 Hz) represented by ALFF could be interpreted as evidence for increased presynaptic synchrony, which would likely enhance spike synchrony albeit at much longer time intervals than previously studied or over much larger neuronal populations than that reflected by spike correlations between neuron pairs. SS has also been found to increase gamma-band (50–100 Hz) oscillatory activity in ACx (Stolzberg et al., 2013); oscillations substantially higher than in ALFF. An alternative view is that the tinnitus percept is derived from coordinated activity among several auditory and nonauditory regions (Horwitz and Braun, 2004; Husain et al., 2006). Enhanced FC between the HIP and auditory areas provides a substrate for assigning a spatial location to a phantom sound, while coordinated activity between specific auditory areas and the reticular formation and AMY may draw attention to and add emotional significance to neural activity in the auditory pathway. Thus, functionally coordinated activity within the network may be essential for bringing tinnitus into consciousness. Reallocating network resources Tinnitus and hyperacusis, like phantom limb pain and cutaneous allodynia, are triggered by peripheral damage presumably leading to widespread changes in the CNS that involve altered connections in networks that include portions of the central auditory pathway and other regions linked to emotion, memory, attention, and arousal (Llinas et al., 1999; Leaver et al., 2012; Husain and Schmidt, 2014). In the resting state, SS increased ALFF and FC in a broad-neural network that included core auditory structures extending from the IC to the ACx consistent with previous studies implicating these central auditory structures in tinnitus (Paul et al., 2009). SS also enhanced sound-evoked LFP in the MGB and ACx suggesting a key role for these auditory structures in amplifying auditory information that could manifest as loudness intolerance (Gu et al., 2010). Cerebellar gating and gain Although the cerebellum is mainly involved in motor planning and control, some cerebellar regions such as the PFL and vermis receive inputs from auditory centers (Petacchi et al., 2005) and respond to sound (Lockwood et al., 1999). Interestingly, the perception of tinnitus has been linked to activation of the PFL and vermis (Brozoski et al., 2007a) consistent with our results. Since ablation or inactivation of the PFL eliminates the perception of noise-induced tinnitus (Bauer et al., 2013), some have suggested that the PFL acts as a gain control mechanism comparing
Cerebral palsy is the most common motor disability of childhood. Spastic cerebral palsy accounts for 60% to 70% of cases. Research has shown that acupuncture can improve the quality of life of children with cerebral palsy, but the mechanism of action remains unclear. This study aims to determine the effectiveness of acupuncture for treatment of children with spastic cerebral palsy and to assess the value of multimodal magnetic resonance imaging (MRI) and ambulatory electroencephalogram (EEG) for evaluation of treatment effect.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)