Figure Press PLAY to activate the animation. Then, click on the cochlea and text to gain further information. This is a fast acting system. These fibers synapse in the dorsal cochlear nucleus, and may function as a general warning as when you might jump from a loud sound.
These fibers decussate and ascend in the lateral lemniscus to the inferior colliculus. This slow acting system involves much more processing and may provide more detailed information about the sound, such as its location. These fibers synapse in the ventral cochlear nucleus. Fibers from the ventral cochlear nucleus synapse in the ipsilateral and contralateral superior olivary nucleus.
Some fibers from the ventral cochlear nucleus cross the midline in the trapezoid body. Thus, cells in the superior olive receive inputs from both ears and are the first place in the central auditory system where binaural processing stereo hearing is possible. The optical tract is a major afferent of this nucleus. The SC also receives projections from the parabigeminal nucleus, substantia nigra pars reticulata, zona incerta and corticotectal projections from visual cortical areas and the medial temporal cortex Olszewski and Baxter, ; May, The SC plays a key role for generation of saccadic eye movements, as it projects to the paramedian pontine reticular formation through the descending predorsal bundle; meanwhile its ascending branch projects toward the interstitial nucleus of Cajal, a vertical saccade generator in the midbrain, and to the mediodorsal thalamic nucleus and the intralaminar thalamic nuclei.
The SOC contains two principal nuclei: the medial superior olivary nucleus and the lateral superior olivary nucleus. The SOC is located in the ventrolateral border of the caudal pontine tegmentum. The function of this nucleus is to process the binaural input converging from the cochlear nuclei from the two ears. The MG is the thalamic relay of the auditory pathway.
It is located in the ventromedial thalamus, it receives input from the IC and projects toward the auditory cortex. Finally, the LG is the thalamic relay of the visual pathway and is located in the ventrolateral thalamus. The LG receives input form the SC and projects toward the visual cortex. Additionally, previous studies performed automatic segmentation of the LG in an attempt to compare human healthy controls to glaucoma patients using T 1 -weighted images acquired with 3 Tesla and 1.
Nevertheless, these studies mainly focused on single-subject nuclei segmentations rather than creating in vivo human probabilistic i. Both auditory and visual pathways demonstrated lower gray matter volumes in disease states relative to healthy controls specifically, lower volume of the right IC and left hippocampus in tinnitus patients Landgrebe et al.
Similarly, other studies suggest that visual deficits can be predicted by delineation of the optic tract de Blank et al. An improved understanding of the anatomical boundaries of these nuclei may better elucidate such prognostic frameworks. Further, acoustic pathway delineation is limited by current MRI techniques Maffei et al.
Finally, LG is relevant in psychiatric diseases like schizophrenia Mai et al. Data were obtained in a prior study Bianciardi et al. While the acquisition and analysis are summarized below, full details can be found in Bianciardi et al. A custom-built channel receive coil and volume transmit coil was employed in data acquisition Keil et al.
The improved sensitivity of this coil was due to: i a more posterior arrangement of the coil elements; ii its shape curving around the back of the head enabling a closer proximity to the cerebellum and brainstem as opposed to commercial coils that typically extend straight down along the lower parts of the head; iii a more efficient flip angle calibration for lower parts of the brain because the coil extends more inferiorly than traditional coils.
We utilized a common single-shot 2D echo-planar imaging scheme for 1. This provided T 2 -weighted anatomical images with resolution and geometric distortions perfectly matched to the DTI dataset.
With the echo-planar-imaging scheme, we were also able to mitigate specific-absorption-rate limits of spin-warp T 2 -weighted MRI at 7 Tesla. Importantly, use of unipolar Stejskal and Tanner, rather than bipolar Reese et al. This template was utilized because it encompasses the entirety of the brainstem, is compatible with diffusion-based tractography, and provides high contrast detail in the brainstem.
The high-dimensional non-linear warp transformation employed a symmetric diffeomorphic normalization transformation model with smoothing sigmas: 3, 2, 1, 0 voxels — fixed image space —, and histogram image matching prior to registration.
It also employed a cross correlation metric, regular sampling, gradient step size: 0. The combined transformation was then applied to both single-subject FA and T 2 -weighted images, using a single-interpolation step interpolation method: linear.
On a single-subject basis, two raters C. Each rater inspected both imaging modalities simultaneously. The manual segmentation was performed by utilizing image contrast and anatomical landmarks Paxinos et al. The IC was delineated based on a hypointense region in T 2 -weighted images located in the tectum of the midbrain, caudal to the SC on the dorsal aspect of the mesencephalon, lateral —in its dorsal aspect- to the periaqueductal gray hyperintense in T 2 -weighted images , and —ventrally— to the cuneiform nucleus hypointense in the FA map.
The IC was bounded laterally by the cerebrospinal fluid. The SC was visible in the midbrain as a hypointense region on T 2 -weighted images, rostral to the IC, lateral to the periaqueductal gray and bounded laterally by the cerebrospinal fluid.
The SOC, comprising the lateral and medial superior olive, the lateroventral and medioventral periolivary nucleus, and the superior paraolivary nucleus, was identified on FA maps as a hypointense oval region dorsolateral to the central tegmental tract, and with its most caudal part posterior to the superior border of the inferior olive.
For each subject and each final label coregistered to single-subject native space via the inverse of the transformations described in 2. We then computed the mean s. For both i and ii , the modified Hausdorff distance Dubuisson and Jain, , a measure of spatial overlap frequently used in neuroimaging Fischl et al. We also evaluated the inter-rater agreement and the internal consistency of labels with a metric more commonly used in the literature especially for larger brain structures , the Dice similarity coefficient Dice, The Dice similarity coefficient of two labels X and Y e.
Finally, for each nucleus, the modified Hausdorff distance as well as the Dice similarity coefficient in both cases i and ii was averaged across subjects. Again, the left and right nuclei appeared as hypointensities on T 2 -weighted MRI, possibly indicating a higher iron concentration compared to neighboring areas Drayer et al.
Figure 1. Very good i. Figure 2. Figure 3. Figure 4. Very good spatial i. Figure 5. The inter-rater label agreement and the internal consistency of each label computed for validation are both shown in Figure 6. For each nucleus, the average modified Hausdorff distance assessing the inter-rater agreement and the internal consistency of nuclei atlas labels Figure 6 , upper row was below the linear spatial imaging resolution 1.
In addition, the Dice similarity coefficient for the inter-rater agreement and the internal consistency Figure 6 , bottom row was always above 0. Figure 6. Atlas validation. Dice similarity coefficient across 12 subjects. We mostly included studies performed in younger human adults [age between 19 and 47 years, except for two studies on older subjects Nara et al.
New tools are needed to fill this gap and enable accurate structural delineations of the underlying deep brain nuclei in future clinical and research studies. Then, for each nucleus, we discuss the segmentation process, in relation to the MRI contrast used for the nucleus delineation, the identification of specific nuclei borders, and the obtained volumes in comparison to literature values.
Further, we discuss the possible impact of the atlas for clinical and research studies. Finally, we acknowledge the limitations of this work and propose possible future extensions. In this work we showed that the use of advanced ultra-high-field MRI sequences allowed in vivo structural imaging of tiny brainstem and thalamic nuclei with high contrast, sensitivity and good spatial resolution with respect to conventional e.
Notably, the 7 Tesla sequence was optimized to: i achieve minimum echo time e. Specifically, these 7 Tesla structural MRI T 2 -weighted and FA maps techniques enabled the single-subject segmentation of subcortical auditory and visual subcortical nuclei based on their contrast with respect to neighboring areas. An important consideration regards the multi-contrast, resolution- and distortion-matched structural image acquisition used in this work.
The latter allowed us to discriminate the structural boundaries of some nuclei due to the improved visualization of multiple contrasts in a common space, and to the use of complementary information derived from different image contrasts.
This extends previous reports Li et al. Crucially, our work also demonstrated the feasibility of generating a validated in vivo stereotaxic probabilistic atlas of these structures after precise coregistration to MNI or another stereotactic space.
This atlas complements existing in vivo neuroimaging atlases of other brain structures Tzourio-Mazoyer et al. They were hypointense oval shaped structures easily identifiable on a T 2 -weighted MRI. The periaqueductal gray and the cuneiform nucleus were located at the anteromedial border of the IC with good contrast in FA maps and the cerebrospinal fluid limited its lateral edge.
For the SC, its inferior edge was clearly located near the most rostral part of the IC, yet its rostro-ventral edge —in proximity to the thalamus— was less clearly defined due to the absence of an abrupt change in contrast in that area. The periaqueductal gray neighbored the medial border of the SC, while the cerebrospinal fluid delimited the posterolateral border of this structure. As visible in Table 1 , our IC volumes were within literature values specifically, closer to the inferior range of the latter , mainly derived from histology and structural MRI studies.
Instead, the volume of the SC was slightly lower than reported literature values. Further validation work with MRI and histology might better elucidate the relationship between the achieved contrast in in vivo T 2 -weighted images and the distribution of iron or of other microstructural properties within different layers of this nucleus Drayer et al.
For the manual delineation of the thalamic auditory MG nucleus we mainly used the T 2 -weighted image contrast. The MG was visible as a hypointense area neighboring at its upper edge the posteroventral thalamic nuclei, bounded inferiorly by the cerebrospinal fluid and limited by the pulvinar on its posterolateral border.
Its medial border was clearly visible in FA maps, due to the high contrast with neighboring white matter bundles, such as the medial lemniscus and the spinothalamic tract. It is interesting to note the high variability in the MG volume values reported in the literature see Table 1 for this structure. For instance, fMRI studies reported larger MG volumes compared to histology and structural MRI studies, possibly due to partial volume effects and artifacts due to draining veins in this region Sitek et al.
The MG volumes obtained in this study were within the range of previous literature reports see Table 1. The MG label was useful to delimitate the posteromedial border of the LG, yet for the other borders of the LG we also used the contrast available in the FA maps.
Notably, we used the FA maps to identify white matter bundles neighboring the LG, such as the optic tract anterior to LG and, in its medial aspect, the mesencephalic peduncles. Interestingly, recent in vivo studies report changes in the volume and shape of LG compatible with the location of the magnocellular layer in dyslexia according to previous hypothesis for this disease Giraldo-Chica et al.
Previous reports of the LG volume show high degree of variability up to two-fold between individuals even when using the same techniques Andrews et al. The LG volumes obtained in the present study were within the range of literature values see Table 1. In the current study we found that the right LG was slightly larger compared to the left LG, in agreement with previous reports Li et al. The SOC was a round-oval hypointense area in FA maps with well-defined borders, yet at its most postero-lateral aspects its boundaries were more difficult to delineate due to its close proximity with other structures e.
St-Amant, M. Medial geniculate body. Reference article, Radiopaedia. Central Nervous System. Medial geniculate complex Corpus geniculatum internum Medial geniculate nucleus. URL of Article. Hariqbal Singh, Parvez Sheik. Thomas P. Naidich, Henri M. Duvernoy, Bradley N. To understand how the MGB projection to the auditory striatum contribute to auditory decision-making, we used chemogenetic tools to assess the effects of the MGB projection inhibition on performance of an auditory frequency-discrimination task.
Upon demonstrating the behavioral relevance of this connection, we used in vivo tetrode recordings to characterize striatal sound responses to the stimuli used in the behavioral task, and then optogenetically dissected the MGB and primary ACx contributions to the striatal sound representation.
Our results indicate that the projections from the MGB and primary ACx differentially modulate striatal auditory information and that these effects are essential for making an auditory frequency-discrimination decision. Right panel, Example images of thalamic projections green and cortical projection red to the auditory striatum.
Images are taken at the auditory striatum. Left panel, schematic diagram of viral injection and CNO infusion setup. Right upper panel, example image of labeled thalamostriatal fibers. Right panel, examples of auditory stimuli. Error bars are s. The curve is fitted with logistic sigmoid function. Red: CNO sessions; Black: saline sessions. Effects of CNO-mediated thalamic inhibition on the slopes of psychometric function e , the numbers of non-reported trials per session f , the number of completed trials per session g , and the reaction time h.
Five weeks after viral infection, we observed strong mCherry signals of the MGB projection in the auditory striatum Fig. We then tested the inhibition effect of this projection mediated by the hM4Di receptor. We applied the synthesized activator, clozapine-N-oxide CNO , into the bath solution after establishing whole-cell patch recordings on striatal neurons.
To test any potential side effects due to nonspecific binding of CNO to other endogenous receptors 23 , we conducted the same experiment on control slices prepared from mice expressing only mCherry, but not hM4Di. The ex vivo efficacy of hM4Di-mediated inhibition of the MGB projection to the auditory striatum suggested that a similar strategy may work in vivo. In brief, a freely moving mouse was placed in a sound-proof chamber as previously described Each trial is self-initiated by the mouse poking its nose in the center port to trigger a sound stimulus, and the mouse learns to associate the frequency of pure tones high versus low with actions going to the left versus the right port for a water reward through trial-and-error Fig.
We then bilaterally implanted cannulas into the auditory striatum for local CNO or saline infusion Fig. We continued to train the mice on task after they recovered from the implantation surgery around 2 weeks , and initiated chemogenetic manipulation to inhibit the MGB projection while the mice performed the task.
In hM4Di-mCherry mice, we observed that the CNO infusion substantially decreased choice accuracy in task performance as compared to saline infusion sessions Fig. To further clarify how inhibiting MGB striatal projections negatively influenced task performance, we performed the following analyses: we first quantified trials in which the sound stimuli were triggered but the mice did not choose either of the side ports.
We did not observe differences in the numbers of these nonreport trials across control or CNO sessions Fig. Next, we counted the number of completed trials per session and observed no differences among control or CNO sessions Fig.
Furthermore, there were no change in reaction times response latency across difficulties Fig. These analyses confirmed that the CNO application did not affect the motivational state or motor capabilities of our experimental mice.
We applied the same strategy to examine the behavioral impacts of the corticostriatal pathway from the primary ACx to the auditory striatum. Our analysis showed that inhibition of this cortical projection also significantly decreased choice accuracy in task performance, but had no effects on reaction time Supplementary Fig.
The findings that both MGB and primary ACx projections to the auditory striatum are critical for auditory decision-making led us to ask what type of auditory information these two projections carry. To answer this question, we first characterized the striatal responses to sound stimuli.
We therefore used pure tones for this test. We implanted tetrodes into the left auditory striatum to record tone-evoked responses from striatal neurons.
These three types can be readily distinguished by their spike waveforms from tetrode recordings Fig. In our tetrode recordings, This closely resembled the proportions of recordings reported in previous studies 26 , Striatal neurons respond to tonal stimuli. Scale bars: 0. Averaged tone-evoked firing rates are plotted against the corresponding tone frequencies.
Solid lines: mean values; shaded areas: standard errors. All three types of striatal neurons displayed diverse responses to pure tones, with some neurons exhibiting rapid onset of transient responses while others showed sustained responses, offset responses or suppression responses Fig. The majority of cells in all three cell types manifested transient onset responses, followed by sustained and suppression responses Fig.
Striatal neurons responded to tone stimuli over a wide range of frequencies Fig. We only included tuned neurons in later tuning analysis. We found that neurons with onset and sustained responses displayed best frequencies to the pure tones over the full range of tested stimuli Fig. Together, our profiling revealed that all three major types of striatal neurons responded to tonal stimuli, and showed frequency preferences.
We next asked from where the auditory striatum receives the sound information. As shown in the Fig. Since the primary ACx also receives ascending projections from the MGB 28 , 29 , 30 , we asked whether the MGB neurons that project to the ACx and striatum are from the same or segregated populations Supplementary Fig. Specifically, as shown in Supplementary Fig. Seven days after the injection, we collected brain tissue and analyzed the images.
Interestingly, several previous studies have reported that neurons in the MGd are broadly tuned to tonal stimuli, while MGv neurons are sharply tuned 30 , 31 , 32 , Moreover, the tonotopic organization is preserved in the MGv 34 , the primary ACx 30 , 35 , 36 , and the primary ACx projections to the striatum These studies, together with our anatomical tracing, suggest that the two types of projections may likely relay distinct acoustic information to the auditory striatum in regulating striatal-dependent auditory decisions.
To test this hypothesis, we selectively and transiently silenced the MGB projection and recorded the striatal sound responses. We then implanted a microdrive with integrated optic fibers and tetrodes into the auditory striatum for sound response recording and thalamic projection silencing Fig.
Similar to the experimental design in Fig. Tone: 0—0. Frequencies are aligned to the best frequency of each single unit and scaled in octave away from the best frequency. The best frequencies are determined separately from control and light-on trials for each single unit. Firing rates at different frequencies of each single unit are normalized to the firing rate at the best frequency from control trials. Left panel: the fitted curves from the same example neurons shown in b and c.
Right panel: the fitted curves from the population data shown in d. Black, control condition; red, light-on condition; gray, offset the black curve to the baseline of the red curve; orange, scaled down the gray curve.
After sorting the recordings based on cell type, we focused on analyzing the responses of MSNs in this study, as they are the most common neuron type, and the only projection neurons in the dorsal striatum We compared striatal single-unit responses to sound stimuli between trials with and without light stimulation i.
Silencing thalamic projections, however, did not alter the best frequency of individual striatal single units, or the tuning width Fig. We further analyzed the tone-evoked responses by fitting the tuning curves with a Gaussian function, and found that the fitted tuning curves in control condition could be transformed to those in light-on condition by subtracting a small offset 0.
These findings suggested that MGB projection to the auditory striatum provided a gain controlling function to the tone-evoked responses. To determine the sound information that is relayed by the primary ACx projections to the auditory striatum, we performed the same tests shown in Fig. In brief, we expressed ArchT in primary ACx neurons via AAV injection, and implanted a microdrive with integrated optic fibers and tetrodes into the auditory striatum for recording and cortical projection silencing Fig.
We performed similar analyses as described above for the MGB pathway studies. We found that silencing the primary auditory cortical projection also substantially decreased the overall amplitude of tone-evoked responses of MSNs Fig. Interestingly, unlike silencing of the MGB projection—which led to a decrease to broad frequencies in responsiveness Fig. Together, these results suggest that the primary ACx projection provides tone-tuned information to the auditory striatum.
Silencing the primary ACx projection to the auditory striatum reduced striatal MSN tonal responses to their best frequencies. The finding that MGB and primary ACx projections carry differential sound information motivated us to extend our analysis to test whether it would be sufficient to block sound response by silencing both projections.
As shown in Supplementary Figure 1 , the MGB and the primary ACx are the two primary auditory areas projecting to the auditory striatum. To test this possibility, we expressed the neuronal silencer ArchT in both MGB and the primary ACx, followed by implanting tetrode bundles with an optic fiber into the auditory striatum to record tone-evoked responses.
We performed similar tests as described in Figs. We found that optical silencing of both MGB and primary ACx terminals largely abolished the striatal sound responses Fig. This indicates that the MGB and the primary ACx are the two major sources of acoustic information to the auditory striatum.
Together with those results from the selective inhibition of either MGB Fig. MGB and primary ACx projections are the two major sources for striatal tonal response.
In this study, using an established auditory frequency-discrimination task, we found that chemogenetic inhibition of the MGB or the primary ACx projection to the auditory striatum impaired task performance.
0コメント