Iba1+ Microglia Exhibit Morphological Differences between Inferior Colliculus Sub-Regions and Their Abutments onto GAD67+ Somata Reveal Two Novel Sub-types of GABAergic Neuron

Microglia have classically been viewed as the endogenous phagocytes of the brain, however, emerging evidence suggests roles for microglia in the healthy, mature nervous system. We know little of the contribution microglia make to ongoing processing in sensory systems. To explore Iba1+ microglial diversity, we employed the inferior colliculi (IC) as model nuclei, as they are characterized by sub-regions specialized for differing aspects of auditory processing. We conducted fluorescent multi-channel immunohistochemistry and confocal microscopy in guinea pigs of both sexes and discovered that the density and morphology of Iba1+ labelling varied between parenchymal sub-regions of IC, while GFAP+ labelling of astrocytes was confined to the glia limitans externa and peri-vascular regions. The density of Iba1+ microglia somata was similar across sub-regions, however a greater amount of labelling was found in dorsal cortex than ventral central nucleus or lateral cortex. To further understand these differences between sub-regions in IC, Sholl and skeleton analyses of individual microglia revealed a greater number of branching ramifications in dorsal cortex. We also quantified abutments of Iba1+ microglial processes onto GAD67+ (putative GABAergic) somata. Cluster analyses revealed two novel sub-types of GAD67+ neuron, which can be distinguished solely based on the quantity of axo-somatic Iba1+ abutments they receive. These data demonstrate Iba1+ microglia exhibit different morphologies and interactions with GAD67+ neurons in distinct sub-regions of the mature, healthy IC. Taken together, these findings suggest significant heterogeneity amongst microglia in the auditory system, possibly related to the ongoing functional demands of their niche.


Introduction
Inhibition is an essential element of neural processing and a defining component of sensory systems. GABAergic inhibition is prevalent in the auditory system, particularly in the principal auditory midbrain nuclei, the inferior colliculi (IC). Around a quarter of neurons in IC are GABAergic (Merchán et al., 2005), which may be why the IC is the most metabolically active nucleus in the mammalian brain (Sokoloff et al., 1977). Understanding how inhibitory cell types vary in different brain regions, to specialize for distinct functions, is a key area of neuroscientific study (Freund and Buzsaki, 1996;Tremblay et al., 2016). Most investigations into sub-types of inhibitory neurons naturally focus on the cells per se, including their morphology, electrophysiological firing characteristics, expression of cytoplasmic calcium binding proteins and RNA transcriptome. Another approach is to characterize and classify GABAergic neurons based on differences in the afferent axosomatic inputs they receive (Ito et al., 2009;Beebe et al., 2016).
The IC has a tonotopic topography that can be divided into distinct sub-regions. The central nucleus (CNIC) is dominated by neurons sharply tuned to simple auditory stimuli.
The dorsal cortex (DCIC) has much broader frequency tuning and receives extensive corticofugal input and is specialized for synaptic plasticity (Herbert et al., 1991;Winer et al., 1998;Bajo and Moore, 2005;Bajo et al., 2010). The other major sub-region is the lateral cortex (LCIC) which exhibits polysensory tuning (Aitkin et al., 1978). Despite the essential role of the IC in hearing, little is known of how glial cells contribute to processing therein.
Microglia are an integral cell type present throughout the brain. The morphology of microglia varies throughout the brain, suggesting adaptation to their surrounding milieu (Lawson et al., 1990). Furthermore, microglia interact with neurons during 'normal' processing and can sense and respond to local chemical signaling (Pocock and Kettenmann, 2007;Wake et al., 2009;Schafer et al., 2012), but these processes remain poorly understood. glycol, 1% polyvinyl pyrrolidone-40 in 0.1M PBS) and stored at -20°C until use (Watson et al., 1986;Olthof et al., 2019).
Mouse anti-GFAP (1:500; monoclonal; clone G-A-5; G3893; lot# 045M4889V; Sigma; RRID: AB_477010) -according to the manufacturer, this antibody is raised against an epitope from the C-terminus of GFAP in purified pig spinal cord (Latov et al., 1979;Debus et al., 1983). The antibody has been shown to recognize a single band of approximately 50kDa and reacts with homologous, conserved residues across mammals (Lorenz et al., 2005). The use of this antibody has been demonstrated in many species, including mouse (Komitova et al., 2005), rat (Lennerz et al., 2008;Sanchez et al., 2009), tree shrew (Knabe et al., 2008), guinea pig (Kelleher et al., 2011;Kelleher et al., 2013) and human (Toro et al., 2006). Labelling observed in this study was consistent with these studies and the known morphology of astrocytes.
Rabbit anti-calbindin D-28k (1:1,000; polyclonal; AB1778; lot# 2895780; Millipore; RRID: AB_2068336) -according to the manufacturer, this antibody recognizes a single band at 28kDa in human, mouse, and rat brain tissues. It does not bind to calretinin and pre-adsorbtion of diluted antiserum with calbindin removed all labelling in human brain (Huynh et al., 2000). Previous labelling of mouse olfactory bulb (Kotani et al., 2010), rat piriform cortex (Gavrilovici et al., 2010) and guinea pig enteric nervous system (Liu et al., 2005) all showed highly selective cytoplasmic labelling of neurons. We observed labelling consistent with previous reports.

Fluorescence immunohistochemistry
Sections through the superior colliculus and the rostral-most third along the rostrocaudal axis through the IC were first used to optimize labelling. Data are presented from sections in the middle third of the IC along the rostro-caudal axis, which contained the CNIC, DCIC and LCIC. The location of each section through the rostro-caudal axis was referenced to an atlas of the guinea pig brainstem (Voitenko and Marlinsky, 1993).
All steps in the labelling protocol involved continuous gentle agitation of sections.
Free-floating sections were brought to room temperature and washed 3x5mins in PBS.
Sections were blocked and permeabilized in 5% normal goat serum (Vector) and 0.05% Triton X-100 (Sigma) in PBS for one hour. Following blocking, a cocktail of primary antibodies was added to the blocking solution and applied to sections overnight at room temperature. The next day, sections were washed 3x5mins in PBS and incubated for two hours in appropriate secondary antibodies (Invitrogen; 1:250 in blocking solution). For double labelling of Iba1 and GAD67, goat anti-rabbit AlexaFluor 488 and goat anti-mouse AlexaFluor 568 were used. For double labelling of calbindin and GFAP, goat anti-rabbit AlexaFluor 488 and goat anti-mouse AlexaFluor 647 were used. For triple labelling of Iba1, GSL1 (pre-conjugated rhodamine fluorophore) and GFAP, goat anti-rabbit AlexaFluor 488 and goat anti-mouse AlexaFluor 647 were used. Sections were then mounted on slides and coverslipped using Vectashield (Vector Labs, H-1000) and kept at 4°C until imaged. All experiments had control slides where the primary, secondary or both the primary and secondary antibodies were excluded. This allowed detection of autofluoresence and any aspecific signal and ensured only labelling from primary and secondary binding to epitope targets was imaged.

Image acquisition
Sequentially acquired micrographs were taken with a confocal microscope (Leica SP5) using a wide field stage and zoom function. Images were acquired via a 40x objective (NA=1.25) for images of the entire cross-section of the IC, and a 63x objective (NA=1.4) for region of interest (ROI) panoramas. Whole IC images were taken using 5µm equidistant slices in the Z-plane to produce maximum intensity tiled projections (pixel size; x & y=0.7583µm, z=50-60µm). For GAD67 and Iba1 ROI panoramas, 5-row x 6-column (432x552µm) tiled images were taken using 1µm z-slices and rendered as maximum intensity projections (pixel size; x & y=0.2406µm, z=40µm).

Image analyses
For Iba1+ cell density estimates, tiled panorama images of the IC were subject to manual cell counts. The peripheral borders of the IC were delineated and a contour drawn, and each image cropped to its respective contour. To make fair comparisons between cases, 450µm 2 grids were placed across each IC panorama image and centered on the middle pixels of each micrograph in ImageJ. Only those grids which were filled entirely by stained parenchymal tissue were subject to counts. Comparisons were then made between cases, such that only grids that were present in images from all four animals were included in calculation of group means and standard deviations per grid.
Maximum intensity projection ROI panoramas were analyzed for i) cell counts, ii) percentage field of view covered analyses, iii) individual Iba1+ cell Sholl analyses, iv) skeleton analyses and v) Iba1 abutting GAD67 analyses using Fiji ImageJ (Abràmoff et al., 2004). For analyses i-iv, panorama micrographs were first processed by filtering monochrome images using a median pixel (1.5) filter and then thresholded to binary by implementing the IsoData algorithm. Cell counts, and percentage field of view covered analyses were then performed using the Analyze Particles plugin. For Sholl analyses, individual Iba1+ microglia were cropped and a series of equidistant radiating 1µm concentric circles were plotted from the center of the cell body to the furthest radiating extent of ramification. Each intersection with a concentric ring was measured. The Skeletonize algorithm was used to display a one-pixel thick framework of each microglial cell. The Analyze Skeleton plugin calculated number of branches and branch lengths for each cell.

0
Five GAD67+ cells were selected randomly from each sub region panorama, with a selection criterion that the entirety of the cell must be contained within the x, y and z dimensions of the field of view. A z-stack was collected beyond the limits of each cell using 1µm slices. The absence or presence of the soma in each slice was determined and these data were used to calculate soma diameter to the nearest µm. Soma perimeters were manually contoured in each slice of each GAD67+ cell. Iba1+ labelling that came into contact with the contour with no pixels between was counted as an Iba1+ process abutting a GAD67+ cell soma. These data were calculated for each Iba1+ cell as well as each GAD67+ cell. The length of Iba1+ labelling abutting the perimeter contour around each GAD67+ cell soma was measured.

Statistical analysis
Data were collected in Excel spreadsheets. Statistical hypothesis testing was performed in Prism 7 (GraphPad). Factorial analyses were conducted using the nonparametric Kruskall-Wallis ANOVA with sub-region as the factor in all cases. Where appropriate, post-hoc tests with Dunn's method were conducted. For post-hoc analyses the α was Šidák corrected for multiple comparisons. Spearman's rank correlations were used to investigate potential associations between dependent variables.
Principal component and two-step cluster analyses were conducted in SPSS v25 (IBM). The two-step cluster analysis employed Euclidean distance measures with Schwarz's Bayesian clustering criterion and classified data into one of the two identified clusters. All 160 cell ROIs were successfully classified by this analysis. A Chi-squared test was used to analyze the ratio of cells in IC sub-regions in each cluster. All reported p values are exact and two tailed.

GFAP+ astrocytes and Iba1+ microglia form the glia limitans externa and neurovascular unit in IC
We first sought to identify the distribution of GFAP+ astrocytes and Iba1+ microglia in adult guinea pig IC. Coronal, 60µm sections showed pronounced GFAP+ and Iba1+ labelling of the glia limitans externa lining the dorsal and lateral borders of the IC ( Figure   1A). Extensive labelling was also distributed medially, lining the cerebral aqueduct, with ramified GFAP+ astrocytic processes radiating into the periaqueductal grey, as well as the commissure of the IC. Interestingly, we found no GFAP+ astrocytes throughout the IC parenchyma, save for sparse labelling of cells in the outermost layers of the DCIC and LCIC.
Conversely, ramified Iba1+ microglia tiled the parenchyma in non-overlapping domains with similar density throughout the IC, as quantified in Figure 1B.
Combining Iba1+ and GFAP+ labelling with the fluorescent-conjugated lectin GSL1 revealed extensive peri-vascular labelling along putative penetrating arteries and arterioles ( Figure 1C). Neurons expressing cytoplasmic calbindin or calretinin were distributed in the outermost regions of the cortices of the IC, matching previous reports (Zettel et al., 1997;Ouda et al., 2012) and in close proximity to vessels and GFAP+ processes ( Figure 1D).
These findings demonstrate that many aspects of IC glial organization mirror those reported in other brain regions, with both GFAP+ astrocytes and Iba1+ microglia forming the glia limitans externa and lining adjacent to blood vessels. However, the observation that Iba1+ microglia but not GFAP+ astrocytes were found throughout the parenchyma, suggests a role for Iba1+ microglia in glial-neuronal putative interactions in IC.

Iba1+ microglia in DCIC are more ramified than other sub-regions of IC
We predicted that the stronger Iba1+ microglia labelling in DCIC neuropil was primarily due to a greater number and extent of ramifications compared to other sub-regions of IC. To test this, we conducted Sholl analyses for a total of 64 cells per sub-region (n=256).
The maximum intensity projection of each Iba1+ microglial cell was imaged and analyzed in x and y dimensions.
Cells were identified and selected from ROI images ( Figure 4C&D).
Background/non-cellular labelling was cropped ( Figure 4E) and cellular labelling thresholded to generate binary images ( Figure 4F). The number of intersections at every micrometer distance from the center of the soma was calculated ( Figure 4G). Binary thresholded cells were also skeletonized to derive information about the shape and structure of ramifications, such as the number of branches and maximum branch length ( Figure 4H). Conversely, maximum branch length, defined as the longest distance covered by any ramification of skeletonized Iba1+ microglia without branching, followed the opposite trend.
Longest maximum branch lengths were found in VCNIC (median=14.44µm; IQR=12.76-17.18) ( Figure 5D). Shorter distances were found in LCIC ( Figure 6. These features were calculated from micrographs such as the representative example in Figure 7A, which shows a GAD67+ neuron being abutted by two Iba1+ microglia. A correlation matrix revealed weak associations between GAD67+ soma maximum diameter (i) and the other four dependent variables (ii-v) ( Figure 7B). There were stronger correlations between the four Iba1 related variables (ii-v). As these variables were only weakly correlated with GAD67+ neuron diameter, we further investigated whether a multivariate analysis could better explain the observed distributions.
We conducted a principal component analysis for the five variables in all sub-regions of the IC. The data showed a clear dissociation between GAD67+ neuron diameter in one cluster and the other four variables, which clustered together ( Figure 7C). Both clusters were categorized using a standard correlation coefficient of >0.5 as a cut-off value, which showed one cluster was explained by only the GAD67+ neuron diameter variable, while the other cluster had significant contributions from all four of the Iba1+ related variables. These trends were also true for all sub-region specific analyses in IC (Figure 7Di-iv). We then conducted a two-step cluster analysis including all five variables. We employed Euclidean distance measures with Schwarz's Bayesian clustering criterion ( Figure 7E). Contrasting iterations up to 15 clusters, the analysis found two clusters demonstrated the best explanatory power, displaying good (0.5) silhouette measures of cohesion and separation ( Figure 7F).
All 160 cell ROIs were classified into one of the two clusters determined by the twostep analysis. Representative examples of each cluster found in each sub-region are shown in Figure 8. There were 23 cases in cluster 1, and 137 in cluster 2. To visualize the contribution of each of the four Iba1 related variables, each was plotted as a function of GAD67+ neuron diameter ( Figure 9A-D). These scatterplots revealed a dissociation with little overlap between the two clusters using the percentage Iba1+ abutting GAD67+ somata ( Figure 9A). There was almost perfect discrimination between the clusters of the normalized total µms of Iba1+ abutments onto GAD67+ neuron somata ( Figure 9B). Conversely, the number of Iba1+ cells abutting each GAD67+ soma had a significant degree of overlap with little difference between clusters ( Figure 9C). The number of Iba1+ processes abutting GAD67+ somata had little overlap between distributions ( Figure 9D).
To compare the ability of each of these variables to independently discriminate between the two clusters, we conducted ROC analyses ( Figure 9E). These data revealed that while each variable had an area under the curve >0.5, the three variables relating to the nature of Iba1+ processes abutting GAD67+ neurons had the best discriminatory power.

Iba1+ putative interactions with GAD67+ neurons show little difference between sub-regions of IC
We explored whether any of the variables or clusters identified had a relationship to the sub-region of IC in which the cells were located. Diameter of GAD67+ somata did not vary between sub-regions ( Figure

Discussion
These findings reveal that Iba1+ microglia, but not those of GFAP+ astrocytes, are present throughout the parenchyma of the IC in the healthy, mature auditory system. Taking advantage of the specialized and functionally diverse sub-regions of IC, we found the first evidence, to our knowledge, revealing differences in microglial morphology, between subregions. Interestingly, Iba1+ microglia in DCIC, a sub-region known to receive a greater proportion of glutamatergic corticofugal contacts, are more ramified than those in other subregions of IC. We also developed a new analysis method, which was used to investigate the number and length of abutting microglial processes onto GAD67+ neuronal somata.
Multivariate cluster analyses were applied to semi-quantify variables of abutments and revealed two distinct types of GAD67+ neuron in IC, distinguished by the extent of Iba1+ microglial abutments onto their somata. Taken together, these data demonstrate that microglia in IC exhibit anatomical features and connectivity with GABAergic neurons suggesting specialization of function relating to the local demands of processing.

Significance of IC sub-regional differences in microglial morphology
Sensory processing can be interpreted through triad models of organization, based on observations that central sub-regions of sensory pathways are dominated by ascending innervation, producing brisk responses at short latencies to simple stimuli, such as in the CNIC. Across different sensory nuclei, there are at least two sub-regions located adjacent to central sub-regions, which typically require more complex stimuli to elicit responses and which occur at longer response latencies. One of these sub-regions typically receives a  (2010)). The present study provides evidence that Iba1+ microglia morphologies also exhibit differences between sub-regions ( Figure 5).
Why might microglia exhibit differing morphologies in distinct sub-regions of IC?
There is longstanding evidence that microglia are sensitive to their local environment and exhibit morphological differences throughout the brain (Lawson et al., 1990). There are clear neuroanatomical differences between CNIC, DCIC and LCIC, including increased cytoarchitectural and myeloarchitectural density in CNIC (Faye-Lund and Osen, 1985). There is also a more defined laminar organization of the CNIC than the DCIC, in part due to the neuronal morphologies present as well as coursing fibres (Oliver and Morest, 1984). As the DCIC is known to have larger neuron sizes than CNIC, one might speculate that the increased number of branching ramifications we observed in DCIC ( Figure 5) correlates with larger neuronal somata. However, the LCIC is also known to have larger neurons than CNIC, but had a similar number of microglial branching ramifications to mid-CNIC and VCNIC. This suggests that differing microglial morphologies may relate to other aspects of local processing. For instance, it is known that microglia interact at synapses ( Microglial influence over these various aspects of auditory processing remains to be explored. Another explanation may be the higher density of neurons in CNIC. There has been previous suggestion that microglial cell density and neuronal cell density are inversely related (Lawson et al., 1990). However, we found no strong evidence for differences in microglial cell density between IC sub-regions using two different levels of approach ( Figure 1B&4A).
This was in spite of much greater numbers of GAD67+ somata in VCNIC than other subregions ( Figure 2&3). A third possibility is that there are relationships with the myeloarchitectural differences between sub-regions, however, DCIC and VCNIC exhibit similar, high levels of axons of passage compared to mid-CNIC, arguing against such an explanation. We therefore have no strong evidence as to why microglia in DCIC are more ramified that other sub-regions and this requires further investigation.

Two novel clusters of GABAergic neurons
As the auditory pathway contains a large proportion of inhibitory neurons, with around a quarter of neurons being GABAergic (Oliver et al., 1994;Merchán et al., 2005), understanding their structure, function, and organization is a question of fundamental 1 importance. Previous approaches to classifying GAD67+ neurons in IC have focused on soma size (Roberts andRibak, 1987a, 1987b;Ono et al., 2005)  It may be claimed that cluster analyses, as well as other classification approaches, can produce artificial discriminations between data that reside along continua, such as has been reported for frequency response areas in IC (Palmer et al., 2013). Indeed in many cases, forcing data through cluster analyses will produce clustered data, irrespective of whether these clusters represent meaningful differences. However, we permuted our cluster analysis through various iterations and found that not only could the data be best clustered by two clusters, but that those clusters had strong explanatory power, with good silhouette measures of cohesion and separation and good cluster quality ( Figure 7E&F). Furthermore, visualization of representative examples of these clusters showed that those cells in cluster 1 clearly received a greater number and length of abutting Iba1+ processes onto their somata than those in cluster 2 ( Figure 8).
ROC analyses revealed that the two clusters could be distinguished by three variables that quantified aspects of Iba1+ processes, but not by the number of Iba1+ microglial cells abutting each GAD67+ neuron. This may reflect the highly motile and dynamic nature of microglial processes (Wake et al., 2009). Other features of GABAergic neurons in IC, such as their discharge patterns and expression of associated ion channels also do not relate to soma size (Ono et al., 2005). Interestingly, the two identified clusters of GAD67+ neurons did not differ in their relative proportion between the four sub-regions in IC. Future work may investigate the differing afferent neural inputs to and efferent targets of these cells, to identify likely physiological and connectional differences between clusters and their relationship to GABAergic processing.

Technical Considerations
The use of primary antibodies in less studied species such as guinea pig can be challenging due to potential differences in epitopes and when not adequately controlled for, may lead to spurious observations (Schonbrunn, 2014). This is important when using exploratory approaches as in the present study, to ensure all analyses are predicated on specific and selective labelling (Voskuil, 2017). We therefore conducted extensive control experiments, excluding primary antibody only, secondary antibody only and both antibodies, to ensure analyses were based on true labelling.
The lack of GFAP+ astrocytes in IC parenchyma (Figure 1) was surprising and necessitated confirmatory experiments. However, in all cases, a lack of GFAP+ astrocytes in the parenchyma was found alongside extensive labelling in peri-vascular regions and the glia limitans externa, demonstrating consistency within and between cases. The lack of GFAP+ astrocytes in IC parenchyma does not exclude the possibility that astrocytes reside throughout IC. Indeed, a recent report employing SR101 revealed a network of putative astrocytes throughout CNIC (Ghirardini et al., 2018). However, there is some labelling of oligodendrocytes with this marker, which hampers interpretability in studies trying to selectively label astrocytes (Hill and Grutzendler, 2014). There are a diversity of other non-GFAP markers that may reveal the distribution of distinct astrocyte sub-types throughout the IC, including the parenchyma, however, this was beyond the scope of the present study.
Functional differences between astrocytes in CNIC and the outer layers of DCIC and LCIC have been suggested previously via 3-chloropropanediol-induced lesions, which selectively destroyed the former but not the latter (Willis et al., 2003;Willis et al., 2004). The present study leads to the speculation of fundamental gliochemical and physiological differences that may relate to the sub-region specific roles microglia and astrocytes play in their local milieux (Lawson et al., 1990;Olah et al., 2011). Recently, RT-PCR of single IC astrocytes revealed expression of functional inhibitory neurotransmitter transporters GlyT1, GAT-1, and GAT-3 (Ghirardini et al., 2018). Sub-regional differences in GAD67+ neurons in the present study suggest that GABAergic and glycinergic signaling released from and received by glial cells may also exhibit such variations throughout IC and perhaps in other structures.

Conclusions
We have described, for the first time, that Iba1+ microglia, but not GFAP+ astrocytes tile the adult IC parenchyma and have discovered sub-regional differences in the morphology of microglia in IC. Furthermore, multivariate statistical approaches revealed two new clusters of GAD67+ neurons which can be distinguished based on the total amount of Iba1+ abutments they receive from microglial processes. Our findings demonstrate morphological (and suggest putative functional) diversity amongst IC microglia, with differential ability to interact with GAD67+ somata. These data highlight the fundamental role microglia play in the organization and likely function of sensory systems in the healthy, mature brain.  (  2  0  1  3  )  I  n  t  r  a  c  e  l  l  u  l  a  r  r  e  s  p  o  n  s  e  s  t  o  f  r  e  q  u  e  n  c  y  m  o  d  u  l  a  t  e  d  t  o  n  e  s  i  n  t  h  e  d  o  r  s  a  l  c  o  r  t  e  x  o  f  t  h  e  m  o  u  s  e  i  n  f  e  r  i  o  r  c  o  l  l  i  c  u  l  u  s  .  F  r  o  n  t  i  e  r  s  i  n  N  e  u  r  a  l  C  i  r  c  u  i  t  s  7  :  7  .  G  h  i  r  a  r  d  i  n  i  E  ,  W  a  d  l  e  S  L  ,  A  u  g  u  s  t  i  n  V  ,  B  e  c  k  e  r  J  ,  B  r  i  l  l  S  ,  H  a  m  m  e  r  i  c  h  J  ,  S  e  i  f  e  r  t  G  ,  S  t  e  p  h  a  n  J  (  2  0  1  8  )  E  x  p  r  e  s  s  i  o  n  o  f  f  u  n  c  t  i  o  n  a  l  i  n  h  i  b  i  t  o  r  y  n  e  u  r  o  t  r  a  n  s  m  i  t  t  e  r  t  r  a  n  s  p  o  r  t  e  r  s  G  l  y  T  1  ,  G  A  T  -1  ,  a  n  d  G  A  T  -3  b  y  a  s  t  r  o  c  y  t  e  s  o  f  i  n  f  e  r  i  o  r  c  o  l  l  i  c  u  l  u  s  a  n  d  h  i  p  p  o  c  a  m  p  u e  r  g  i  c  n  e  u  r  o  n  s  a  n  d  a  x  o  n  t  e  r  m  i  n  a  l  s  i  n  t  h  e  b  r  a  i  n  s  t  e  m  a  u  d  i  t  o  r  y  n  u  c  l  e  i  o  f  t  h  e  g  e  r  b  i  l  .  J  o  u  r  n  a  l  o  f  C  o  m  p  a  r  a  t  i  v  e  N  e  u  r  o  l  o  g  y  2  5  8  :  2  6  7  -2  8  0  .  R  o  b  e  r  t  s  R  C  ,  R  i  b  a  k  C  E  (  1  9  8  7  b  )  A  n  e  l  e  c  t  r  o  n  m  i  c  r  o  s  c  o  p  i  c  s  t  u  d  y  o  f  G  A  B  A  e  r  g  i  c  n  e  u  r  o  n  s  a  n  d  t  e  r  m  i  n  a  l  s  i  n  t  h  e  c  e  n  t  r  a  l  n  u  c  l  e  u  s  o  f  t  h  e  i  n  f  e  r  i  o  r  c  o  l  l  i  c  u  l  u  s  o  f  t  h  e  r  a  t  .  J  o  u  r  n  a  l  o  f  n  e  u  r  o  c  y  t  o  l  o  g  y  1  6  :  3  3  3  -3  4 W  a  t  s  o  n  R  E  ,  W  i  e  g  a  n  d  S  J  ,  C  l  o  u  g  h  R  W  ,  H  o  f  f  m  a  n  G  E  (  1  9  8  6  )  U  s  e  o  f  c  r  y  o  p  r  o  t  e  c  t  a  n  t  t  o  m  a  i  n  t  a  i  n  l  o  n  g  t  e  r  m  p  e  p  t  i  d  e  i  m  m  u  n  o  r  e  a  c  t  i  v  i  t  y  a  n  d  t  i  s  s  u  e  m  o  r  p  h  o  l  o  g  y  .  P  e  p  t  i  d  e  s  7  :  1  5  5  -1  5  9  .  W  i  l  l  i  s  C  L  ,  L  e  a  c  h  L  ,  C  l  a  r  k  e  G  J  ,  N  o  l  a  n  C  C  ,  R  a  y  D  E  (  2  0  0  4  )  R  e  v  e  r  s  i  b  l  e  d  i  s  r  u  p  t  i  o  n  o  f  t  i  g  h  t  j  u  n  c  t  i  o  n  c  o  m  p  l  e  x  e  s  i  n  t  h  e  r  a  t  b  l  o  o  d  -b  r  a  i  n  b  a  r  r  i  e  r  ,  f  o  l  l  o  w  i  n  g  t  r  a  n  s  i  t  o  r  y  f  o  c  a  l  a  s  t  r  o  c  y  t  e  l  o  s  s  .  G  l  i  a  4  8 :       separation were compared from 2-15 clusters, with two clusters demonstrating the highest cluster quality. These data demonstrate that not only did two clusters exhibit the best available explanatory power, but also strong clustering quality, reflecting real underlying differences in Iba1+ abutments onto GAD67+ neurons between clusters. ROC analyses showing classifier performance of each variable in discriminating GAD67+ cells into cluster 1 or cluster 2. Normalized total µm Iba1+ abutting and number of processes of GAD67+ somata could almost perfectly discriminate between clusters.