These studies uncover two downstream signaling pathways defined by a kinase (AAK1) and a GEF (Rabin8), which regulate complex neuronal dendritic and synaptic phenotypes orchestrated by NDR1/2. NDR1 and NDR2 transcripts

have been found in the brain by reverse transcription polymerase chain reaction (RT-PCR) and northern blot (Devroe et al., 2004 and Stegert et al., 2004), and NDR2 mRNA has been localized via in situ buy PD-0332991 hybridization in various brain regions, including the hippocampus and cortex (Stork et al., 2004). To determine the developmental profile of NDR1 and NDR2 expression, we probed brain lysates from postnatal day (P)5, P10, P15 and P20 via a mouse monoclonal antibody raised against NDR1 and a polyclonal antibody we generated that is specific for NDR2 (see Experimental Procedures). Both antibodies recognized a major protein band, which was present throughout development, at ∼55 KD (Figures 1A; Figure S1A available online). The NDR1 antibody did not recognize overexpressed NDR2, and the NDR2 antibody did not recognize overexpressed NDR1 in COS-7 cells, demonstrating their specificity (Figure S1B). Immunocytochemistry using these antibodies revealed that NDR1 and NDR2 are present in the cytoplasm

in hippocampal pyramidal neurons and in the cortex (Figure 1B and data not shown) and are found throughout the cell body and dendrites in dissociated hippocampal neurons in culture (Figure 1C). NDR1 was also present in the nucleus in agreement with previous reports (Millward et al., 1999; data not shown). In order to investigate NDR1/2′s cell autonomous Adriamycin in vitro function in dendrite development, we used three approaches. Dominant-negative or constitutively active NDR1/2 expression, siRNA knockdown of NDR1 and NDR2, and a chemical genetics approach to block NDR1 activity were used. NDR1 mutations used in this study are shown in Figures 1D and 1E. We found similar results with all three approaches. The biochemical activation mechanism of NDR kinases has been established. MST3 kinase phosphorylates NDR1/2 at its C-terminal hydrophobic

residue T444 to activate Phosphatidylinositol diacylglycerol-lyase it (Stegert et al., 2005). NDR1/2 can be activated by okadaic acid (OA) via inhibition of protein phosphatase 2A, facilitating phosphorylation at T444 and the autophosphorylation at S281 (Stegert et al., 2005). MOB1/2 binding to the N-terminal region of NDR kinases is required for the release of auto-inhibition and maximal activity (Bichsel et al., 2004). Autophosphorylation site S281 is critical for NDR1/2 kinase activity. In order to test NDR1/2′s role in dendrite development, we first generated dominant negative and constitutively active NDR1 mutants (Figures 1D and 1E). For dominant negative NDR1, we mutated Ser281 and Thr444 to Alanine (S281A; T444A, NDR1-AA) or catalytic lysine to alanine (K118A, NDR1-KD); both mutants have no kinase activity (Millward et al., 1999 and Stegert et al., 2004).

This further supports SLC6A15 as the gene of interest within this locus. The reduction of SLC6A15 expression in CA1 could be validated by in situ hybridization in nine stress-susceptible versus nine Alectinib chemical structure stress-resilient mice ( Figures 6A and 6B). Moreover, a significant reduction in SLC6A15 expression could also be observed in the dentate gyrus of stress susceptible animals ( Figures 6C and 6D). The demonstrated downregulation of SLC6A15 expression in stress-susceptible mice, most prominent in the CA1 region of the hippocampus, suggests SLC6A15 to play a role in long-term effects of chronic stress on neuronal circuits and is in accordance with the human MD risk genotype-dependent

effects, assayed by in vivo volumetry, which were also strongest in the CA subregion of the hippocampus. We performed a GWA study, with replication of the top hit and genome-wide significant association in the meta-analysis across a total of 4,088 patients and 11,001 controls, including one sample from a different ethnic background. Together with gene expression data, neuroimaging correlates and evidence

from a mouse model of chronic stress our results point toward SLC6A15, a neuronal amino acid transporter, Epacadostat chemical structure as a candidate gene in the pathophysiology of major depression. Even though the direction of the association of rs1545843 with depression and depressive symptoms was the same in all samples with nongeriatric depression, the effect sizes were heterogeneous, with a much larger effect in the discovery sample (OR = 2.8 for the recessive model) as compared to the other samples with odds ratios ranging

from 1.18 to 1.61. The strong association and low p value in the discovery sample is probably due to the “winners’ curse,” but this phenomenon has also been observed below for other, now established, disease loci. For example, the association of a SNP in the FGFR2 gene with breast cancer was much stronger in the rather small discovery sample than in any of the subsequent replication samples. However, the direction of the association was consistent and reached a p value of 2e-76 in close to 30,000 cases and controls ( Easton et al., 2007). This indicates that heterogeneous effect sizes with overestimation of the effect in a small discovery sample may still be in agreement with a true signal. In addition to the genome-wide significance ( Dudbridge and Gusnanto, 2008) observed in our study, replication of the effect in samples of different ethnicities, European and African-American, might be a further indicator for a true effect. In addition to replication in independent samples, the functional relevance of the associated locus is supported by results of gene expression analyses in premortem human hippocampus and EBV-transformed lymphoblastoid cell lines of the HapMap individuals (Stranger et al., 2005) and peripheral blood monocytes (Heinzen et al., 2008).

For example, mutations in transporters that confer susceptibility to blockade by exogenous small molecules that have no effects on native

proteins could allow acute and reversible inhibition of transporters in astrocytes. It is worth considering some of the physical and chemical constraints screening assay for functional hyperemia. First and foremost, since neurovascular coupling is spatially confined, the molecular signals need to be generated and communicated locally. Therefore, it is important to determine the range of integration and range of influence of astrocytes, especially since astrocytes are extensively coupled through gap junctions (Haydon, 2001). Second, even if the vasoactive signals are generated locally, they may spread far if they diffuse rapidly and have a long lifetime—these LY2109761 research buy parameters need to be measured for molecules such as NO and ions such as potassium. Third, affecting blood vessels in one place may affect the perfusion nonlocally because of vascular connectivity and passive redistribution of blood (Boas et al., 2008). These considerations have been recognized for some time but are not

always attended to in molecular and cellular studies. Finally, the same messenger might have different or even opposing effects on blood flow (Attwell et al., 2010). Another open issue is related to the spatial “reach” of astrocytes. Cortical astrocytes are organized into nonoverlapping functional domains (Halassa et al., 2007) (Figure 2C). On the input side, a single

cortical astrocyte can, in principle, listen to tens of thousands of synapses by virtue of its extensive processes (Haydon, 2001), but it is unclear how many synapses are MycoClean Mycoplasma Removal Kit needed to activate an astrocyte. Recent in vivo experiments in visual cortex indicate that astrocytes respond to visual stimulation with calcium rises with exquisite selectivity, suggesting that their “input” field may be highly selective (Schummers et al., 2008). Selective astrocytic responses were also found in slice experiments in barrel cortex (Schipke et al., 2008). On the other hand, what is the spatial extent of a single astrocyte’s output? In theory, the organization of astrocytes into separate domains may contribute to the spatial distribution of the CBF response (Iadecola and Nedergaard, 2007). However, the input and output selectivity may not be limited by the spatial extent of a single astrocyte’s processes, since extensive gap junction coupling of astrocytes may extend the range substantially by allowing intercellular transfer of signaling molecules (Haydon, 2001 and Scemes and Giaume, 2006). The degree of astrocyte coupling may also be regulated to make network topology modifiable. The extent of astrocyte coupling in vivo is unclear. One signaling event that has been observed to propagate across astrocytes is a rise in calcium concentration. Calcium waves spreading across multiple astrocytes were imaged more than two decades ago in vitro (Cornell-Bell et al.

At P15, the difference in expression was even more dramatic ( Figures 6C and 6D), with many fewer neurons expressing GFP in ThGVdKO mice, and these cells were largely distributed deeper in cortex, corresponding to L5a, than in control mice and did not express CUX1 ( Figure 6D). Thalamocortical axon terminal selleck chemicals llc arbors at P15 completely overlapped with the layer of neurons

expressing GFP ( Figure 6E), consistent with the dominant expression of Dcdc2a-Gfp in L4 of control mice. In contrast, in ThVGdKO mice, GFP neurons were present mainly below the bulk of thalamocortical axon terminal arbors ( Figure 6E). These data suggest that the normal maintenance in L4 and downregulation in L5a of Dcdc2a expression

are disrupted in ThVGdKO mice, possibly due to disruptions in postnatal neuronal position or changes in laminar expression of the Dcdc2a-Gfp reporter. We examined the expression of a number of genes with layer-specific expression patterns in ThVGdKO somatosensory cortex at P15 and consistently observed changes in and around L4. As already described, the expression of the predominantly superficial layer gene Cux1 in ThVGdKO mice was significantly reduced ( Figures 3G–3K), as was SatB2 ( Figures S5A and S5B). The expression of the L4 transcription factor Rorb (RORβ; Schaeren-Wiemers et al., 1997), and the L5a transcription factor Etv1 (aka Er81; Yoneshima et al., 2006), changed reciprocally

( Figures 7A–7D). In ThVGdKO, a dense band of Rorb-positive cells that corresponds to L4 was MDV3100 in vivo shifted upward ( Figures 7A and 7B), consistent with the Nissl staining ( Figures 3C and 3D). We also Oxalosuccinic acid observed a number of Rorb-positive cells throughout L5 and L6 in ThVGdKO mice, which was more unusual in controls ( Figure 7A, black arrows; Figures S6A and S6B, white arrows). In control mice, neurons expressing Rorb were mostly confined to L4 and coexpressed CUX1, whereas in ThVGdKO mice, Rorb expression extended to L5 where it was not coexpressed with CUX1 ( Figure S7). The domain of Etv1 expression spread toward the pial surface in ThVGdKO mice ( Figures 7C and 7D), again consistent with the expansion of L5 observed with Nissl staining. The expression of L5b (Ctip2, Fezf2) and L6 (FoxP2, Tbr1) markers was largely undisturbed in the somatosensory cortex of ThVGdKO mice ( Figures S5C–S5G, and data not shown), and changes in laminar-specific gene expression observed in somatosensory cortex did not occur in the motor cortex of ThVGdKO mice ( Figures S6C and S6D). None of these aberrant expression patterns were apparent at P6 ( Figure S6E). The changes in layer-specific gene expression observed in ThVGdKO cortex at P15 are consistent with the differences observed histologically and imply an unexpected degree of activity-dependent thalamic influence on laminar development of somatosensory cortex.

, 1999 and Ishikane et al., 2005). Those studies had suggested that these cells detect large approaching objects through synchronized oscillatory activity and thereby this website trigger the frog’s escape

response to dark looming objects. It is conceivable that this detection of slowly approaching objects works in concert with the detection of suddenly appearing large objects proposed in the present study, thus mediating detection of large, potentially threatening objects over a wide range of behavioral scenarios by a single ganglion cell type in the amphibian retina. Inhibitory signaling appears to be critical for both mechanisms, and indeed, it was shown that the frog’s natural escape behavior to large objects is suppressed when inhibition is blocked in the frog’s eyes (Ishikane et al., 2005). The circuit structure proposed to generate the local gain control of homogeneity detectors (Figure 7C) is not unique to the retina. Similar parallel transmission pathways of excitation and inhibition have also been identified elsewhere in the brain (Porter et al., 2001, Pouille and Scanziani, 2001, Gabernet et al., 2005, Sun et al., 2006, Cruikshank et al., 2007, Strowbridge, 2009 and Bellavance et al.,

2010). Measuring iso-response stimuli in these systems by stimulating local subcircuits independently could PD0325901 chemical structure be used to probe their functional roles and the potential implementation

PD184352 (CI-1040) of a similar local gain control. Moreover, we suggest that measurements of iso-response stimuli can function as a general tool to identify the rules of signal integration in a wide variety of neural systems wherever different signaling streams converge onto single neurons. Retinas were isolated from dark-adapted adult axolotl salamanders (Ambystoma mexicanum; pigmented wild-type) and mounted onto 60 channel multielectrode arrays for extracellular recordings of ganglion cell spiking activity. All experimental procedures were performed in accordance with institutional guidelines of the Max Planck Society. During the recordings, retinas were superfused with oxygenated Ringer’s solution at room temperature (20°C–22°C). For experiments with pharmacological inhibition block, strychnine (5 μM), picrotoxin (150 μM), and bicuculline (20 μM) were added to the Ringer’s solution. Visual stimuli were displayed by a gamma-corrected cathode ray tube monitor and focused onto the retina with standard optics. Mean light intensities were in the photopic range, and all stated contrast values correspond to Weber contrast (Istimulus – Ibackground) / Ibackground.

A similar

argument can be put forward against the theory that well-timed (phase locked), contralateral inhibition originating from the MNTB delays the time point at which the action potential threshold is reached (Brand et al., 2002; Pecka et al., 2008). This theory provides an elegant explanation for the observation that best ITDs typically show a bias for contralateral lead, which we also observed in the present study. This theory also predicts a significant phase-dependent influence of the sound from one ear on the response to the sound presented to the other ear, since well-timed inhibition should interact with excitation even if it is entirely of the shunting type. see more In contrast to these predictions, we found that the timing of the input from either Cobimetinib solubility dmso ear is unaffected by the phase of the input from the other ear. Our results therefore suggest that the timing

of the inhibitory input from either ear is not sufficiently precise to allow it to shift the ITD tuning (Joris and Yin, 2007; Zhou et al., 2005). This argument still holds true in the presence of inhibition from both ears. We cannot entirely exclude that the use of anesthetics may have influenced the timing precision of the inhibition. Effects of ketamine/xylazine on subcortical auditory processing are typically mild (Smith and Mills, 1989; Ter-Mikaelian et al., 2007), and both bushy cells (Kuenzel et al., 2011) and primary neurons of the MNTB aminophylline (Hermann et al., 2007; Kopp-Scheinpflug et al., 2008) in gerbil show considerable spontaneous activity even under ketamine/xylazine anesthesia. Decreased inhibition has been reported in the dorsal cochlear nucleus (Navawongse and Voigt, 2009). However, the original evidence favoring well-timed inhibition was also obtained under ketamine/xylazine anesthesia (Brand et al., 2002; Pecka et al., 2008). Another possible confounder is that most of the inhibition is somatic and may have been disrupted when we made recordings. However, somatic inhibitory responses

in the MSO are not disrupted by positive pressures at least ten fold higher than what we used during approach of cells for juxtacellular recordings (Couchman et al., 2012). The presence in the MSO of strong glycinergic inhibitory inputs originating from both the ipsi- (LNTB) and contralateral ear (MNTB) is well established, but its function has been debated (reviewed in Grothe et al., 2010). Because of the linearity of the interaction between both ears, a role of well-timed inhibition in shifting the best ITD (Brand et al., 2002; Pecka et al., 2008) seems unlikely. The low variance at the worst ITD suggests that it is the periodic absence of excitatory input rather than phase-locked inhibition that sets the firing rate during the worst ITD. A possible role for inhibition is that it may improve the dynamic range of the MSO neurons, similar to its proposed role in the nucleus laminaris (Yamada et al.

Bodily self-consciousness can be conspicuously modified by pathological and physiological factors. An example of a body-part-specific self-identity disorder is the feeling that one’s own limb does not belong to oneself. These complex misperceptions and misconceptions are comparatively common after cerebral lesions in the right temporo-parietal lobe and typically affect

the left limb (Berlucchi and Aglioti, 2010). Patients with disruptions in full-body self-awareness, generally referred to as autoscopic phenomena (AP), report Gemcitabine in vivo bizarre feelings and exhibit strange behaviors that mimic psychiatric more than neurological disease symptomologies. AP are characterized by the illusory sensation of a second body seen in extracorporeal Selleckchem PD-1/PD-L1 inhibitor space. At least three different forms of AP have been described, namely autoscopic hallucination (the person sees a

second own body with self-location normally anchored to the physical body), heautoscopy (self-location is perceived at both the physical and the illusory body), and out-of-body experience (OBE) characterized by a sense of disembodiment, with the illusory body, to which self-location is attributed, perceived in a position elevated with respect to the physical body (Blanke and Metzinger, 2009). Studies of patients with OBEs suggest that the feeling of being outside the real body and looking at the world from another perspective might be linked to temporo-parietal and vestibulo-insular brain dysfunction (Blanke and Metzinger, 2009). Tellingly, OBEs, as well as the somatosensory feeling that someone else is close to ADAMTS5 us even if nobody is around, have been induced by electrical stimulation of the temporo-parietal regions (Blanke and Metzinger, 2009 and Arzy et al., 2006). Our clear and stable sense of bodily self-consciousness

can also be challenged by simple psychophysical manipulations. For example, touch stimuli delivered to one’s visually obscured or “unseen” hand while observing the synchronous stroking of a seen rubber hand induces the subjective perception that the rubber hand belongs to oneself (rubber hand illusion, Botvinick and Cohen, 1998 and Ehrsson et al., 2004). Inclusion of an inanimate rubber hand into one’s own body representation is not observed when a time lag between visually perceived and physically sensed tactile stimulation is introduced (asynchronous stimulation condition, Botvinick and Cohen, 1998). Using a similar visuo-tactile stimulation paradigm and virtual reality techniques, Lenggenhager et al. (2007) have been able to induce the subjective feeling of whole-body self-identification with an avatar.

Figure S3A shows the green/red fluorescence ratios over time in a single experiment, while Figure S4B shows the average rate of increase in green/red fluorescence over three independent experiments. Fibroblasts carrying VCP mutations exhibit similar lipid peroxidation rates when compared to controls ( Figures S3A–S3C). These results suggest that uncoupling in these cells is not related to changes or alterations in lipid peroxidation Ulixertinib molecular weight rates. Basal cellular ATP levels are determined by the rates of ATP production (oxidative phosphorylation and glycolysis) and

consumption. To monitor ATP levels in live VCP-deficient cells, we used a FRET-based ATP sensor. In a first subset of experiments, control and VCP KD SH-SY5Y cells were treated with 100 μM glycolytic inhibitor iodoacetic acid (IAA) to monitor the ATP levels generated from glycolysis and then with 0.2 μg/ml oligomycin to determine the ATP levels generated by the ATP synthase ( Figures 4A, 4B, and S4A). In a second group of measurements, ATP synthase was inhibited prior to inhibition of glycolysis ( Figure S4B). In all experiments, basal ATP levels were measured prior to treatment with inhibitors. Figure 4A shows traces from a representative experiment in which ATP levels were measured in untransfected, SCR,

and VCP KD cells. The relative ATP levels generated by glycolysis and the ATP synthase as selleck products seen as a reduction in the YFP/CFP ratio after addition of inhibitors are represented in Figure 4B. A not statistically significant decrease in ATP levels was observed in VCP KD cells after inhibition of glycolysis by IAA ( Figures 4B and S4A). However, ATP levels were significantly lower in VCP KD cells compared to controls after inhibition of ATP synthase by oligomycin ( Figure 4B and S4A) (YFP/CFP: untransfected = 0.21 ± 0.07, n = 3; SCR = 0.30 ± 0.05, n = 4; VCP KD = 0.04 ± 0.01, n = 4). Interestingly, when glycolysis was inhibited after ATP synthase, control and VCP KD SH-SY5Y cells showed no decrease in ATP levels in response to IAA ( Figure S4B). Due to the low efficiency of transfection in primary patient

fibroblasts, a bioluminescent assay based on the luciferin-luciferase system was used to detect the ATP levels in these cells. In all three patient fibroblast lines, ATP levels were significantly Resminostat decreased compared to age-matched control fibroblasts (luminescence arbitrary units: patient 1 = 0.63 ± 0.05, n = 7; patient 2 = 0.66 ± 0.07, n = 5; patient 3 = 0.53 ± 0.09, n = 7; control 1 = 0.90 ± 0.09, n = 7; control 2 = 0.91 ± 0.03, n = 6; control 3 = 0.90 ± 0.06, n = 4) ( Figure 4C). These experiments show that VCP pathogenic mutations also lead to decreased ATP levels. The energy capacity was then measured in VCP-deficient fibroblasts to determine the cause of low ATP levels in these cells. The energy capacity of a cell is defined as the time between application of inhibitors of glycolysis/ATP-synthase (i.e., cessation of ATP production) and the time of cell lysis (i.

One example is shown in Figure 6B. Its injection site is shown by a magenta square in Figure 6A. The saccade latency bias, which was present before the

injection (Figure 6B, left), disappeared 10 min after the muscimol injection (Figure 6B, right). We repeated the same experiment (bilateral muscimol injections in the VP) four times, see more and the results were very similar (Table 1). To examine how localized the effects of muscimol injections were, we made bilateral muscimol injections at 12 sites around the VP (Table 1). The reward-dependent saccade latency bias decreased after most injections, but the effects tended to appear later than after the injections in the VP (i.e., longer latencies). We graded the effectiveness of the injection based on the latency (Sakamoto and Hikosaka, 1989): highly effective if the latency was less than 20 min and

weakly effective if the latency was more than 20 min but less than 40 min. Such effective injection sites are depicted by magenta (highly effective) and green (weakly effective) in Figure 6A. As shown in Figure 6A, the highly effective sites were centered on the VP (AC+1), but extended posteriorly into the GPe and GPi, which was 3 mm posterior to the VP (AC-2). At a more anterior level (3 mm anterior to the VP, AC+4) injections into the nucleus accumbens (NAc) had no clear effect selleckchem (n = 2). At a more posterior level (6 mm posterior to the VP, not shown) injections into the subthalamic nucleus (STN)/SNr induced oculomotor effects that prevented the monkey from making saccades to the target, such as involuntary saccades and nystagmic eye movements (Hikosaka and Wurtz, 1985). These data suggest that the effects of muscimol on the reward-dependent saccade latency bias were relatively localized in a region including the VP and possibly the GPe-GPi. Figure 6C shows the population Rolziracetam data based on muscimol injections into the effective sites (n = 11), indicating similarly devastating effects on the reward-dependent saccade latency bias. Crucially, these inactivations selectively affected the motivational

bias, without severe impairments in the sensorimotor control of saccades. If anything, inactivation caused saccade latencies to become shorter, especially on during small-reward trials for on which saccades normally had very long latencies (Figure 6D; Figure S4A). These bilateral muscimol injections also caused changes in saccade peak velocity and the stability of gaze fixation (Figure S4B and S4C). The bias in the saccade peak velocity (i.e., faster when a large reward was expected) disappeared after the injections. After the muscimol injections the monkey broke fixations more frequently before the target came on (error rate of fixation break: p < 0.01 for both small- and large-reward conditions, Wilcoxon signed-rank test). We found that VP neurons encoded the expected value associated with the upcoming action (i.e.

, 1976, Brown et al., 1980b, Bryan et al., 1973, Craig, 1976, Enevoldson and Gordon, 1989b, Hongo et al., 1968, Lundberg, 1964 and Taub and Bishop, 1965). On the basis of fiber and cell body counts, there are an estimated 4,000–6,000 SCT

neurons in the cat, with a much more even spread along the rostrocaudal extent in comparison to PSDC neurons, which seem to be concentrated Rapamycin cell line in cervical and lumbar enlargements. Most SCT neurons are located within lamina IV and have dorsally directed dendrites that terminate abruptly at the lamina II/III border. The majority have cone-shaped dendritic trees, with a few displaying more prominent ventral dendritic arborizations (Figure 4C). Like PSDC neurons, SCT neurons have axon collaterals that extend several segmental levels and may have local actions in spinal reflex pathways (Brown, 1981b). The neural components of the dorsal horn, which include presynaptic sensory inputs, locally projecting interneurons, descending modulatory inputs, and long-range projection neurons, are linked by a highly complex set of synaptic connections. Dorsal horn neurons not only receive synaptic input from primary afferents but also from neighboring excitatory

and inhibitory neurons, each with relative input strengths that most likely differ among modules of neuronal connections. Though our knowledge of dorsal horn circuit organization is still in its selleck compound infancy, recently old gained genetic access

to both pre- and postsynaptic neurons will allow for modality-specific dissection of dorsal horn circuits. As with all primary afferents, LTMRs use glutamate as their principal fast transmitter; therefore, all LTMR subtypes have an excitatory action on their postsynaptic targets of the dorsal horn (Brumovsky et al., 2007 and Todd et al., 2003). However, synaptic arrangements between LTMR subtypes and their postsynaptic targets can be quite complex, often forming synaptic glomeruli, structures that not only include primary afferent axonal boutons and postsynaptic dendrites but also synaptic contacts with axons of neighboring interneurons. The presence of synaptic glomeruli allows for input modulation at the very first synapse within the dorsal horn and is thus thought to be the anatomical substrate for primary afferent presynaptic modulation. Within the dorsal horn, two main types of synaptic glomeruli have been described. Type I glomeruli are present largely in lamina II, have dark primary afferent axons, are thought to arise from unmeylinated fibers and axonal contacts that are GABA reactive, and are thought to arise from purely GABAergic interneurons.