Brain Struct Funct (2013) 218:1279–1292 DOI 10.1007/s00429-012-0457-7
ORIGINAL ARTICLE
Three-dimensional reconstruction and quantitative morphometric analysis of pyramidal and giant neurons of the rat dorsal cochlear nucleus ´ ron K} Szilvia Kecskes • A oszeghy • Ge´za Szu¨cs • • Zolta´n Ruszna´k Clara Matesz • Andra´s Birinyi
Received: 21 June 2012 / Accepted: 14 September 2012 / Published online: 6 October 2012 Springer-Verlag Berlin Heidelberg 2012
Abstract Correct interpretation of functional data obtained from various cell types of the cochlear nucleus (CN), a structure involved in auditory information processing, necessitates reliable cell identification. Our aim was to perform a quantitative morphological characterization of giant and pyramidal cells of the rat CN and identify parameters that are suitable for their adequate classification. Neurons were labeled with biocytin, visualized with a fluorescent marker, and three-dimensionally reconstructed from confocal images. The size and shape of the soma and dendritic tree of each neuron were characterized by 17 morphometric parameters. The variables were subjected to multivariate statistical analysis to determine their importance while discriminating between giant and pyramidal cells. Our results provide a new battery of morphometric data, which could not be obtained earlier, improve the chances of correct cell identification, make modeling experiments easier and more reliable, and help us to understand both the functions of individual CN neurons and the network S. Kecskes A. Birinyi (&) Department of Anatomy, Histology and Embryology, Faculty of Medicine, Medical and Health Science Center, University of Debrecen, Debrecen 4012, Hungary e-mail:
[email protected] ´ . K}oszeghy G. Szu¨cs Z. Ruszna´k A Department of Physiology, Faculty of Medicine, Medical and Health Science Center, University of Debrecen, Debrecen 4012, Hungary
properties of this nucleus. In addition, we demonstrate that even partial labeling and/or incomplete reconstruction of neurons may be enough for their correct identification if selected parameters describing the cell bodies and the proximal portions of the dendritic trees are utilized. We propose that our findings have specific relevance to studies which attempt cell identification after functional experiments resulting in incomplete labeling of the investigated neurons. Keywords Brainstem Fluorescent labeling Morphology Multivariate statistics Abbreviations 3D Three-dimensional ACSF Artificial cerebrospinal fluid CAN Canonical axis CN Cochlear nucleus DCN Dorsal cochlear nucleus HEPES 4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid IC Inferior colliculus Kv Depolarization-activated K? channel L-ACSF Low-sodium artificial cerebrospinal fluid PB Phosphate buffer TBS TRIS-buffered saline VCN Ventral cochlear nucleus VIAS Volume Integration and Alignment System
Z. Ruszna´k Neuroscience Research Australia, Sydney, NSW 2031, Australia
Introduction
C. Matesz Department of Anatomy, Histology and Embryology, Faculty of Medicine, Medical and Health Science Center, HAS-UD Neuroscience Research Group MTA-TKI 355, Hungary, University of Debrecen, Debrecen 4012, Hungary
The mammalian central auditory pathways originate from the cochlear nucleus (CN). Better understanding of the function of this nucleus, including the characterization of the roles of the various cell types it accommodates, would
123
1280
increase our chances of designing better cochlear implants. Ideally, future cochlear implants should produce more ‘natural’-like stimulation of the acoustic nerve and should, therefore, provide higher quality sound perception than the currently available models. However, to explore and characterize the signal processing performed by the individual cell types of the CN, their activity patterns should be correctly associated with the type of neuron they belong to. This necessitates reliable cell identification procedures. In most species, the CN is divided into ventral and dorsal parts (VCN and DCN, respectively; e.g., Osen 1969). Giant and pyramidal (fusiform) cells are considered as the principal cell types of the DCN because they are the only neurons whose axons project out of this nucleus (e.g., Oliver 1984; Ryugo and Willard 1985; Cant and Benson 2003). Giant cells have large, polygonal cell bodies, which emit at least three dendritic trees that extend as far as 500–600 lm from the cell body (Zhang and Oertel 1993; Ostapoff et al. 1994; Alibardi 1999). The longest diameter of the giant cell body ranges from 28 to 70 lm, e.g., mouse (Zhang and Oertel 1993), rat (Alibardi 1999; Malmierca 2003; Pocsai et al. 2007; Pa´l et al. 2009), cat (Kane et al. 1981), and primates (Heiman-Patterson and Strominger 1985). Giant cells project to the ipsi- and contralateral inferior colliculus (IC) and to the contralateral CN (e.g., Cant and Gaston 1982; Shore et al. 1992; Schofield and Cant 1996). The cell bodies of the pyramidal cells are situated in the fusiform layer of the DCN (Blackstad et al. 1984; Hancock and Voigt 2002; Pa´l et al. 2003). The dendritic trees of the pyramidal neurons form two major groups known as apical and basal dendrites (Lorente de No 1981; Webster and Trune 1982; Rhode et al. 1983). Apical dendrites intrude the molecular layer of the DCN and contact the parallel fibers, whereas basal ones enter the deep parts of the DCN and make contact with the acoustic nerve fibers. Similar pyramidal cell morphology has been described in several species, including mouse (Zhang and Oertel 1994), rat (Malmierca 2003), hamster (Schweitzer 1990), gerbil (Hancock and Voigt 2002), cat (Blackstad et al. 1984), guinea pig (Manis 1990), and primates—including humans (Heiman-Patterson and Strominger 1985). Pyramidal cells project to the ipsi- and contralateral IC (Osen 1972; Ryugo et al. 1981). Although giant and pyramidal cells represent two different cell types, they share a number of similarities: (a) when compared to other types of neurons in the CN, they both have large cell bodies and extensive dendritic arborizations; (b) their axons leave the DCN via the dorsal acoustic stria (Smith et al. 2005); (c) they project to the IC; (d) they possess strong calbindin, but lack calretinin-specific immunopositivity (Po´r et al. 2005); (e) their activities are regulated by inhibitory, GABAergic and glycinergic inputs (Davis and Young 2000); (f) they receive information from the auditory nerve fibers and produce a firing
123
Brain Struct Funct (2013) 218:1279–1292
pattern which consists of an initial monosynaptic excitation and a disynaptic inhibition (Babalian et al. 2003); (g) they are not capable of producing ‘primary-like’ response (i.e., a faithful copy of the incoming activity); (h) they are contacted by descending axons originating from the ipsi- and contralateral auditory cortex (Schofield and Coomes 2005); and (i) they are regarded as parts of the ‘monaural pathways’ (e.g., Kudo and Nakamura 1987). Based on the recognition of frequency notches that are introduced by the filtering properties of the pinna, pyramidal cells may send information to the IC about the position of sound source in space (Young et al. 1992). Giant cells may have similar function (Zheng and Voigt 2006). Since no marker has been hitherto reported that would distinguish between giant and pyramidal cells (e.g., Wright et al. 1996; Perney and Kaczmarek 1997; Juiz et al. 2000; Fredrich et al. 2009), their correct identification should rely on their morphology. In one of our previous studies, CN neurons of 1-month-old rats were labeled retrogradely with tetramethylrhodamine-dextran followed by confocal imaging (Pocsai et al. 2007). We have shown that the cell bodies and proximal processes of giant and pyramidal neurons have heterogeneous appearances and correct identification of these cell types may be difficult. Unfortunately, our rhodamine-labeling study was handicapped by the inability of the tracer to penetrate the distal dendrites of the neurons; thus, it could not reveal the entire dendritic tree of the investigated cells. Consequently, this important piece of information could not be taken into consideration while determining cell identity. Although both giant (Friauf 1994; Frisina et al. 1995; Spatz 1997) and pyramidal cells express calbindin (Rogers and Resibois 1992; Frisina et al. 1995; Korada and Schwartz 2000), calbindin-specific labeling is limited to the cell bodies and cannot visualize the whole dendritic arborization either (Po´r et al. 2005). As an alternative attempt, we applied rhodamine backfilling in combination with immunohistochemistry to establish the depolarizationactivated K? channel subunit (Kv) expression patterns of CN neurons (Ruszna´k et al. 2008). However, neither giant nor pyramidal cells showed specific and uniform expression pattern for any of the several Kv subunits tested. In the present work, CN neurons were filled with biocytin and their cell bodies and dendritic trees were threedimensionally reconstructed from confocal images. Cells were identified as giant, pyramidal, or neither. The latter group, referred as ‘unclassified’ hereafter, accommodated several different cell types with one common feature: they did not belong to either the giant or the pyramidal group. Using multivariate statistical methods, our results provide parameters that permit reliable identification of giant and pyramidal cells. We propose that these new parameters are useful for the classification of CN neurons, even if their morphology is not completely revealed.
Brain Struct Funct (2013) 218:1279–1292
Materials and methods Animal care and slice preparation The experiments were conducted on 10- to 14-day-old Wistar rats of both sexes (12 females and 23 males). The experimental protocols were authorized by the Committee of Animal Research of the University of Debrecen and they were in accordance with the relevant international and Hungarian laws. Brain slices were prepared in ice-cold low-sodium artificial cerebrospinal fluid—(L-ACSF) containing (in mM): KCl 2.5, sucrose 250, NaHCO3 26, glucose 10, NaH2PO4 1.25, CaCl2 2, MgCl2 1, and ascorbic acid 0.5—employing a protocol that was published earlier (Ruszna´k et al. 1997; Pa´l et al. 2009). Briefly, animals were killed by decapitation then the brains were removed and transferred into ice-cold L-ACSF. The whole brain was bisected along the midline and the cerebral and cerebellar hemispheres were removed. The two halves of brainstem were glued to a metal block on their medial surface using a cyanoacrylate glue. Parasagittal slices (with a thickness of 200 lm) were cut from the brainstem employing a Microm HM 650 V vibratome (Microm International GmbH, Walldorf, Germany). After preparation, slices were transferred into a temperature controlled and continuously bubbled (95 % O2 and 5 % CO2), chamber, which was filled with normal artificial cerebrospinal fluid (ACSF) having the following composition (in mM): NaCl 125, KCl 2.5, NaHCO3 26, glucose 10, NaH2PO4 1.25, CaCl2 2, MgCl2 1, myo-inositol 3, ascorbic acid 0.5, and Na-pyruvate 2. The slices were kept at 37 C for 30 min, then the chamber was allowed to cool down to room temperature (approx. 22 C). Biocytin labeling Biocytin labeling of single neurons of the DCN was carried out with the application of whole-cell patch-clamp. The brain slices were viewed with a Zeiss Axioskop microscope (Carl Zeiss AG, Oberkochen, Germany) equipped with a 639 water immersion objective and differential interference contrast optics. During the electrophysiological recordings, the slices were continuously perfused with ACSF. Microelectrodes were pulled from borosilicate glass capillaries (Warner Instruments LLC, Hamden, CT, USA) with a Narishige vertical puller (Narishige, Tokyo, Japan). The resistance of the microelectrodes was approximately 2 MX when filled with the internal solution containing (in mM): K-gluconate 114, KCl 4, 4-(2-hydroxyethyl)-1piperazineethanesulfonic acid (HEPES) 10, K2-ATP 4, Na3-GTP 0.3, Na2-phosphocreatinine 10, and biocytin 8. Biocytin was allowed to diffuse into the neurons from the micropipette for 15–20 min. The brain slices were fixed
1281
overnight with 4 % paraformaldehyde dissolved in 0.1 M phosphate buffer (PB; pH = 7.4). The slices were washed with TRIS-buffered saline (TBS) for 10 min. Permeabilization was achieved by incubating the slices in TBS supplemented with 0.1 % Triton X-100 and 10 % bovine serum for 60 min. For the fluorescent visualization of biocytin, the slices were incubated in TBS supplemented with 0.1 % Triton X-100, 1 % bovine serum, and 0.33 % streptavidin-conjugated Alexa488 (Molecular Probes Inc., Eugene, OR, USA) for 90 min. Finally, the sections were washed again in TBS and the slices were mounted with Vectashield mounting medium (Vector Laboratories Inc., Burlingame, CA, USA). Confocal microscopy Biocytin-filled cells were visualized with a Zeiss LSM 510 META confocal laser scanning microscope. Z-stack images were captured from the labeled neurons with an optical thickness of 2 lm through a 409 oil immersion objective. To cover the entire dendritic arborization for the threedimensional (3D) reconstruction, 9–41 series of Z-stack images were acquired from each neuron at different fields in the x–y plane. The separate image stacks of the same neuron were joined using the ‘‘Volume Integration and Alignment System’’ (VIAS) software (Wearne et al. 2005; Computational Neurobiology and Imaging Center, http://www.mssm.edu, Mount Sinai School of Medicine, New York, NY, USA). Reconstruction of intracellularly labeled cochlear nucleus neurons Three-dimensional (3D) reconstructions of 35 labeled CN neurons were performed using the Image Stack Module of the Neurolucida 8.0 program (MBF Bioscience, Inc., Williston, VT, USA). On the basis of their appearances, 14 giant and 16 pyramidal neurons were identified, whereas 5 cells did not fall into either category. The latter cells formed the ‘unclassified’ group. The completeness of all dendritic trees was systematically checked and only neurons having completely or nearly completely visualized arborizations were considered for the quantitative description of their dendritic morphology. Due to incomplete labeling of their dendrites, 12 neurons were excluded from the further quantitative analysis. Consequently, parameters characterizing the dendritic trees were determined on the basis of 9 giant (having 43 dendritic trees), 9 pyramidal (31 dendritic trees), and 5 ‘unclassified’ (23 dendritic trees) neurons. A further scrutiny of the reconstructed dendritic trees revealed that, in some cases, a few processes were truncated during the slicing and/or some of the distal dendritic branches appeared to be insufficiently
123
1282
Brain Struct Funct (2013) 218:1279–1292
labeled. To check for these eventualities, the relationship between the diameter of stem dendrites and the total length of the completely labeled dendritic trees was estimated using linear regression. This procedure permitted the approximation of the length of the dendritic tree for each stem dendrite. Because the difference between the calculated and measured length of the dendritic trees exceeded 20 %, one more giant and two more pyramidal cells were excluded from the further analysis. Consequently, parameters describing entire dendritic arborizations were obtained from eight giant, seven pyramidal, and five ‘unclassified’ neurons. Ten of these cells had only 0–1 incomplete dendritic ending and the proportion of the chopped segments was less than 10 % in all remaining cases indicating that all dendritic trees selected for dendritic analysis were fully contained within the thickness of the sections. Morphometric analysis The quantitative analysis of the morphology of the reconstructed CN neurons was performed with Neurolucida Explorer. Because it is advisable to consider as many parameters as possible for having a useful set of criteria assisting cell classification (Tyner 1975; Rowe and Stone
1977) and because we wanted to increase our chances to successfully discriminate between pyramidal and giant cells, each neuron was characterized by 17 variables (Table 1). The complexity of dendritic trees was estimated using a modified Sholl analysis (Sholl 1953) which was performed in a globe that enclosed the entire dendritic arborization of each investigated cell. This procedure involved defining concentric spheres with gradually increasing radius (with a step size of 10 lm) around the cell body. The number of segments passing to each sphere was calculated and plotted as the function of distance from the common center of the concentric spheres. Statistical analysis The statistical analysis was carried out with the IBM SPSS Statistics software (SPSS Inc., Chicago, IL, USA). In the first step, the individual variables were tested using the Mann–Whitney test, and then the same variables were subjected to multivariate discriminant analysis to test whether the differences detected by univariate statistics are sufficient for unambiguous classification of the pyramidal and giant cells. Discriminant analysis is a technique for classifying objects into predefined classes. The purpose is to determine the class of an observation based on a set of
Table 1 Morphological variables used for the quantitative analysis of the reconstructed neurons Group Soma
Stem dendrite Dendritic tree
123
Morphological variable
Description
Maximum diameter
The distance between the farthest points along the perimeter of the cell body
Surface
The sum of the perimeter and thickness of a series of cylinders describing the cell body
Enclosed volume
Calculated from the surface of the soma with Cavalieri’s method
Roundness
The ratio between the maximum and minimum diameters of the cell body
Convexity
Calculated as ‘convex perimeter/perimeter’ and gives information about the irregularities of the soma
Number
Number of dendrites connected to the soma
Average diameter
The mean diameter of the individual stem dendrites
Sum of diameters
The cumulative diameter of all individual stem dendrites
Diameter of segments Total length
The mean diameter of all dendritic branches The cumulative length of all dendritic segments
Total surface
The cumulative surface of all dendritic segments
Number of endpoints
Number of endpoints of dendritic trees
Average of maximum order
The highest order of each tree was determined according to the centripetal method and the average was calculated for each neuron
Distance to endpoints
The mean distance of endpoints from the soma, as measured along the dendritic branches (path distance)
Convex hull, volume and surface
Approximates the size of the dendritic field. The tips of the most distal dendrites were connected and then the volume and surface area of the convex polygon were calculated
Polar histogram; number of wedge C50 %
The area occupied by the dendritic tree of the neuron was divided into 36, radial, pie-shaped wedges originating from the cell body. The length of the dendritic segments within each wedge was measured, and the section where the length of dendrites reached the maximum was considered as 100 %. The ‘‘number of wedge C50 %’’ variable refers to the number of sections where the lengths of dendrites were more than half of the maximum
Brain Struct Funct (2013) 218:1279–1292
variables known as predictors or input variables. In the first step, each neuron was represented as a point in a multidimensional coordinate system with axes representing the calculated morphometric variables. The analysis transformed this multidimensional space into a two-dimensional canonical plane, where the canonical axes (CAN1, CAN2) were calculated as linear combination of the morphometric variables in the form: L ¼ a1 b1 þ a2 b2 þ þ an bn þ c; where ‘‘a1, a2…an’’ are discriminant coefficients, ‘‘b1, b2…bn’’ are morphometric variables, and ‘‘c’’ is a constant. The orthogonal canonical axes were passed through the multidimensional space in such a way that the separation of the groups was as much as possible relative to the differences among cells within each group. The position of the cells in the canonical plane could be visualized by plotting the individual discriminant scores on the two canonical axes. Finally, the procedure tested the possibility that the discriminant scores of different neuron groups are significantly different and, therefore, the neurons form distinct groups in the canonical space. The correlations between the canonical and original variables were characterized by the Pearson’s coefficients, which were also used to determine the relative importance of the morphological data when discriminating between groups.
1283
3–7 stem dendrites (Figs. 1a, 2a–f) and their dendritic trees divided profusely around the cell body. Some of the dendrites ran toward the surface of the CN and demonstrated rich arborization there (e.g., Fig. 2a, d). The rest of the segments proceeded toward the deeper (i.e., more central) parts of the CN and provided less extensive arborization (Figs. 1a, 2a, d). As opposed to giant cells, neurons identified as pyramidal ones had oval shape or triangular somata situated in the second (fusiform cell) layer of the DCN. The cell bodies of the pyramidal neurons usually gave rise to 3–5 stem dendrites. One of the distinguishing features of the pyramidal cells was the arrangement of their stem dendrites which arose from the opposite poles of the elongated cell bodies (Figs. 1b, 2g–l). These disjunct apical and basal dendritic trees presented long, non-branching segments close to the cell body, and a large number of short branches relatively far away from the soma (Figs. 1b, 2g–l). Although the position of the cell body within the nucleus and the visible parts of the dendritic arborization may assist in cell identification (e.g., Fig. 2a–c, g–i), there are instances when giant and pyramidal cells may be confused with each other (e.g., Fig. 2d–f, j–l). To provide previously unobtainable morphometric data which may increase the chances of adequate classification of labeled neurons, a quantitative analysis of the morphological features of giant and pyramidal cells was performed.
Results Qualitative morphological characteristics of giant and pyramidal cells: initial classification
A quantitative approach for the description of the morphology of giant and pyramidal cells: cell bodies and dendritic trees
Typically, neurons identified as giant cells presented polygonal cell bodies which were located in the deep regions of the DCN. The somata of the giant cells emitted
In our first analysis, 20 completely reconstructed neurons were used (Table 2). Statistical analysis of the individual morphometric variables showed that giant and pyramidal
Fig. 1 Confocal images of biocytin-filled giant (a) and pyramidal (b) cochlear nucleus neurons demonstrating their typical appearance in the sagittal plane. Scale bars 50 lm
123
1284
Brain Struct Funct (2013) 218:1279–1292
Fig. 2 Neurolucida reconstructions illustrating the characteristic features of giant (a–f) and pyramidal (g–l) cells. d–f and j–l 3D reconstructions of the same neurons as shown in Fig. 1. All neurons are presented as they would appear in the sagittal (top row of images), frontal (middle row), and horizontal planes (bottom row), following
Neurolucida reconstruction and shrinkage correction. Dashed lines (in the top row of images) demonstrate the pial surface of the brainstem; black dots mark incomplete endings of the dendritic trees. Scale bars 50 lm
cells differ significantly when the diameter and shape of their cell bodies are considered (for an overview of the distribution of some of the most relevant parameters, see Fig. 3). However, the quantitative analysis of the morphometric parameters indicated that, in contrary to our expectations, the maximum diameter of the pyramidal cell bodies was significantly greater than that of the giant cells. Since this observation raised the possibility that our cell classification was wrong, all parameters (i.e., location and shape of the cell bodies, localization, size, and morphology of the dendritic trees) were checked again, and it was confirmed that all cells had been appropriately identified. Besides the parameters describing the morphology of the cell bodies, there was a statistically significant difference between the number of stem dendrites of giant and pyramidal cells. Both giant and pyramidal cells presented large dendritic arborizations; the parameters that characterized their
dendritic trees (i.e., total length and surface of the dendritic trees, number of endpoints, convex hull) were about twice as large as those associated with other (‘unclassified’) cells. Most parameters characterizing the shape of the dendritic trees of the giant and pyramidal cells were similar and only the polar histograms of their dendritic trees were significantly different. In all of the aforementioned assessments, pair-wise comparisons were made, where only one descriptor was employed at a time. To see whether the differences detected by univariate statistical analysis of the individual variables are enough for unambiguous classification of giant and pyramidal cells, discriminant analysis was employed. Because the neurons available for the analysis were divided into three groups (giant, pyramidal, and ‘unclassified’ cells), two canonical variables explain the total variance among the groups. When all morphological variables were simultaneously taken into consideration
123
Brain Struct Funct (2013) 218:1279–1292
1285
Table 2 Parameters characterizing biocytin-labeled cochlear nucleus neurons having completely preserved morphology Variable
Giant neurons n = 8
Pyramidal neurons n = 7
Unclassified neurons n = 5a
Soma Maximum diameter (lm) 2
41.5 ± 3.82 (30.1–58.7)b
31.3 ± 1.67 (26.7–42.2)
25.3 ± 2.35 (19.9–33.3)
Surface (lm )
1,369 ± 238 (612–2,399)
1,280 ± 142 (884–1,926)
1,036 ± 78 (825–1,271)
Enclosed volume (lm3)
6,092 ± 1,504 (2,448–13,388)
3,885 ± 825 (1,554–7,077)
2,966 ± 519 (1,414–4,311)
Roundness
1.59 ± 0.34 (1.16–2.07)
2.36 ± 0.46 (1.89–2.87)b
1.78 ± 0.75 (1.16–2.81)
Convexity
0.98 ± 0.007 (0.93–1.00)
0.98 ± 0.004 (0.93–1.00)
0.94 ± 0.011 (0.91–0.97)
Number of dendrites
5.13 ± 0.48 (3–7)
3.57 ± 0.37 (3–5)b
4.80 ± 0.66 (3–7)
Average diameter (lm)
2.94 ± 0.31 (1.66–4.47)
3.54 ± 0.30 (2.54–4.40)
3.08 ± 0.32 (2.20–4.10)
Sum of stem diameters (lm)
14.00 ± 1.37 (8.00–21.70)
12.23 ± 0.91 (7.70–15.50)
13.26 ± 1.16 (11.00–17.60)
1.17 ± 0.14 (0.72–1.95)
1.14 ± 0.07 (0.89–1.52)
1.11 ± 0.09 (0.87–1.35)
6,956 ± 276 (5,888–8,315)
7,379 ± 521 (6,242–9,876)
Stem dendrites
Dendritic tree Diameter of segment (lm) Total length (lm) 2
Total surface (lm )
21,463 ± 2,024 (14,672–33,313)
20,459 ± 2,308 (13,600–30,213)
3,936 ± 1,136 (1,599–7,932)c 10,966 ± 2,410 (5,635–18,937)c 77.2 ± 10.5 (51–101)c
Number of endpoints
123.8 ± 13.5 (86–212)
149.1 ± 13.2 (85–184)
Average of maximum order
12.38 ± 1.05 (8–17)
14.57 ± 0.81 (11–17)
13.00 ± 0.95 (10–16)
179 ± 7 (152–214)
132 ± 22 (62–195)
Distance to endpoints (lm)
174 ± 12 (132–249)
Convex hull Volume (lm3) 2
Surface (lm )
368,680 ± 76,928 (154,129–781,704)
290,164 ± 95,909(126,157–722,266)
26,021 ± 2,954 (14,864–35,385)
22,027 ± 3,319 (13,367–37,797)
171,879 ± 78,234 (35,087–443,772) 14,429 ± 4,756 (6,545–31,645)
Polar histogram Number of wedge C50 %
14.63 ± 0.56 (12–17)
11.00 ± 1.40 (4–15)b
11.40 ± 1.44 (6–14)
Values are mean ± SEM. Numbers in brackets show the minimum and maximum values n Number of neurons a
Cells not belonging to either the giant or the pyramidal cell group
b
Significant differences between giant and pyramidal cells (paired comparison, Mann–Whitney test, P \ 0.05)
c
Significant differences between giant/pyramidal and other DCN neurons (P \ 0.05)
(Fig. 4), pyramidal and giant cells formed disjunct clusters along the first canonical axis, and all cells were classified into their original groups (Wilk’s lambda = 0.05, P \ 0.05). In these instances, the first canonical variable was mainly weighed by the size and shape of the soma, and to a lesser extent by the length of the dendritic trees. A quantitative approach for the description of the morphology of giant and pyramidal cells: cell bodies with partially labeled dendritic trees To test whether the previously introduced parameters are applicable for the identification of incompletely reconstructed neurons, the same approach was employed for cells with only partially visualized dendritic trees. To establish whether the parameters characterizing the somata
and stem dendrites may be useful in this scenario, a larger sample of CN neurons was analyzed, including 14 giant, 16 pyramidal, and 5 ‘unclassified’ cells (Table 3). Pair-wise comparisons of pyramidal and giant cells did not show significant differences between the majority of the variables describing the size of the soma. Our test, however, detected significant differences between the maximum diameter (with pyramidal cells having larger cell bodies) and roundness. The larger mean roundness (2.17) associated with pyramidal cells indicates that they have more elongated somata, whereas the cell bodies of giant cells may be better approximated by a polygon (roundness = 1.59). Moreover, the polygonal cell bodies of giant cells emitted 3–7 (mean = 4.86) stem dendrites, whereas the elongated somata of pyramidal cells gave rise to fewer (3–5; mean = 3.56) but significantly thicker stem dendrites
123
1286
Brain Struct Funct (2013) 218:1279–1292
Fig. 3 Distribution of the most characteristic parameters describing the morphology of all reconstructed giant and pyramidal cells. a–d Scatter plots demonstrating the distribution of parameters characterizing completely reconstructed neurons. e, f Scatter plots of parameters characterizing the cell bodies and stem dendrites. g, h Scatter plots of parameters characterizing the dendritic trees. Each symbol represents an individual neuron (a–f) or a dendritic tree (g, h). Squares and triangles indicate giant and pyramidal cells, respectively
(Fig. 3e, f). When cells with partially labeled dendritic trees were considered, the only parameter showing statistically significant difference between the giant/pyramidal and the ‘unclassified’ groups was the maximum diameter of the cell body. When multivariant analysis was applied using the eight variables characterizing the cell bodies and stem dendrites, giant and pyramidal neurons were reasonably well separated along the second canonical axis (Wilk’s lambda = 0.25, P \ 0.001), which correlated with the roundness of the soma and the number of stem dendrites (Fig. 5). The discriminant analysis also indicated that giant and
123
pyramidal cells could be separated from other neurons of the DCN along the first canonical axis, which was mainly influenced by the maximum diameter and the surface of the somata (Wilk’s lambda = 0.53, P \ 0.01). A quantitative approach for the description of the morphology of giant and pyramidal cells: a comparison of the dendritic trees In the last step of the analysis, it was tested whether the qualitative differences noted in the morphologies of the dendritic trees of giant and pyramidal cells can be
Brain Struct Funct (2013) 218:1279–1292
Fig. 4 Analysis of completely reconstructed cochlear nucleus neurons. a Scatter plot showing the distribution of the three neuronal groups on the plane of the canonical axes (CAN1, CAN2) calculated on the basis of all 17 morphological variables. Squares indicate giant cells, triangles correspond to pyramidal cells, and circles show neurons not belonging to either group (‘unclassified’ neurons). b– j Neurolucida reconstructions of the encircled giant (b–d),
1287
pyramidal (e–g), and ‘unclassified’ (h–j) neurons as they would appear in the sagittal (top row), frontal (middle row), and horizontal (bottom row) planes following Neurolucida reconstruction and shrinkage correction. Dashed lines indicate the pial surface of the brainstem; black dots mark incomplete endings of the dendritic trees. Scale bars 50 lm
123
1288
Brain Struct Funct (2013) 218:1279–1292
Table 3 Parameters characterizing the soma and stem dendrites of biocytin-labeled cochlear nucleus neurons Variable
Giant neurons n = 14
Pyramidal neurons n = 16
Unclassified neurons n = 5a
Soma 37.8 ± 2.2 (23.4–58.7)b
25.7 ± 2.3 (19.9–33.3)c
Maximum diameter (lm)
33.7 ± 2.0 (26.7–53.9)
2
1,310 ± 177 (612–2,609)
1,128 ± 91 (616–1,926)
1,036 ± 78 (825–1,271)
5,413 ± 1,075 (1,780–13,388)
3,422 ± 489 (1,265–7,077)
2,966 ± 519 (1,414–4,311)
Surface (lm ) 3
Enclosed volume (lm ) Roundness
1.59 ± 0.25 (1.16–2.07)
2.17 ± 0.33 (1.32–2.87)b
1.56 ± 0.10 (1.16–2.81)
Convexity
0.98 ± 0.007 (0.93–0.99)
0.98 ± 0.004 (0.95–1.00)
0.94 ± 0.011 (0.91–0.97)
Number of dendrites
4.86 ± 0.31 (3–7)
3.56 ± 0.22 (2–5)b
4.80 ± 0.66 (3–7)
Average diameter (lm)
3.05 ± 0.21 (1.66–4.47)
3.84 ± 0.24 (2.52–5.80)b
3.08 ± 0.32 (2.20–4.10)
Sum of diameters (lm)
14.00 ± 0.83 (8.00–21.70)
13.21 ± 0.68 (7.70–20.10)
13.31 ± 1.16 (11.00–17.60)
Stem dendrite
Values are mean ± SEM. Numbers in brackets show the minimum and maximum values n Number of neurons a
Cells not belonging to either the giant or the pyramidal cell group
b
Significant differences between giant and pyramidal cells (paired comparison, Mann–Whitney test, P \ 0.05)
c
Significant difference between giant/pyramidal and other DCN neurons (P \ 0.05)
Fig. 5 Scatter plot showing the distribution of identified (giant and pyramidal) and ‘unclassified’ neurons in the DCN on the plane of the canonical axes (CAN1, CAN2) calculated according to morphological variables characterizing the soma and stem dendrites only. Squares indicate giant cells, triangles correspond to pyramidal cells, and circles show neurons not belonging to either group (‘unclassified’ cells)
quantified. To achieve this, a univariate statistical analysis was applied on 97 labeled and three-dimensionally reconstructed dendritic trees (Table 4). Some of these dendritic trees belonged to cells that have been analyzed in conjunction with Table 2 but the dendritic trees of several other, incompletely labeled neurons were also considered for this phase of the work. Altogether 43 giant, 31 pyramidal, and 23 ‘unclassified’ dendritic trees were analyzed using six variables. The Sholl analysis indicated that individual segments of the dendritic trees of giant and pyramidal cells are distributed differently (Fig. 6). While the dendritic trees of giant neurons profusely branch in the proximity of the cell body, providing a larger number of dendritic segments here, the apical and basal dendritic trees
123
of pyramidal cells extend farther away from the soma and present more complex branching patterns (as indicated by the larger number of endpoints and maximum order; see also Fig. 3g, h). However, most likely owing to the smaller diameters of the pyramidal segments, the total surface area available for synaptic connections did not show significant differences between giant and pyramidal cells. The detailed quantitative analysis of the dendritic arborization also demonstrated that both giant and pyramidal cells possessed significantly larger and more complex dendritic trees than cells belonging to the ‘unclassified’ group. Despite the obvious dissimilarities revealed by classical statistical methods, the multivariant statistical analysis indicated that the dendritic trees belonging to giant, pyramidal, and ‘unclassified’ CN neurons formed three heavily overlapping clusters in the canonical space (Fig. 7). It seems, therefore, that giant and pyramidal cells cannot be reliably distinguished from one another when only parameters characterizing their dendritic arborizations are utilized.
Discussion In the present work, 3D reconstruction and a detailed morphometric analysis of biocytin-labeled pyramidal and giant neurons of the rat DCN were performed. Our aim was to identify quantitative morphometric parameters that may ensure reliable cell classification. Application of multivariate statistical methods justified that giant and pyramidal neurons can be completely separated from each other and from other CN neurons if quantitative features characterizing their somato-dendritic morphologies are compared.
Brain Struct Funct (2013) 218:1279–1292
1289
Table 4 Parameters characterizing the dendritic trees of biocytin-labeled cochlear nucleus neurons Variable
Giant neurons n = 43
Diameter of segments (lm) Total length (lm) 2
Total surface (lm )
Pyramidal neurons n = 31
Unclassified neurons n = 23a
1.15 ± 0.07 (0.57–2.54)
1.05 ± 0.05 (0.64–1.75)
1.12 ± 0.06 (0.72–1.64)
1,451 ± 116 (210–3,400)
2,087 ± 203 (419–5,130)b
854 ± 179 (56–3,712)c
4,518 ± 415 (588–11,184)
5,967 ± 638 (1,015–13,888)
2,378 ± 432 (236–8,772)c
Number of endpoints
25.8 ± 2.3 (3–68)
40.0 ± 4.4 (4–114)b
16.7 ± 2.3 (2–41)
Average of maximum order
9.45 ± 0.47 (3–17)
11.61 ± 0.58 (4–17)b
8.61 ± 0.78 (2–16)
Distance to endpoints (lm)
181 ± 6 (69–267)
186 ± 9 (96–342)
129 ± 14 (22–261)c
Values are mean ± SEM. Numbers in brackets show the minimum and maximum values n Number of neurons a
Cells not belonging to either the giant or the pyramidal cell group
b
Significant differences between giant and pyramidal cells (paired comparison, Mann–Whitney test, P \ 0.05)
c
Significant differences between giant/pyramidal and other DCN neurons (P \ 0.05)
Quantitative morphometric analysis of pyramidal and giant neurons: comparison with earlier data
Fig. 6 Distribution of the dendritic segments of giant and pyramidal neurons on the basis of Sholl analysis. Squares and triangles indicate giant and pyramidal cells, respectively
Fig. 7 Scatter plot showing the distribution of the three investigated neuronal groups on the plane of the canonical axes (CAN1, CAN2) calculated according to morphological variables associated with the dendritic trees. Squares indicate giant cells, triangles correspond to pyramidal cells, and circles show neurons not belonging to either group (‘unclassified’ cells)
The technique applied in the present work may allow reliable cell identification even following microelectrode experiments resulting in incompletely labeled neurons.
We applied multivariate statistical analysis for the identification of pyramidal and giant neurons of the DCN. This technique has been successfully employed for the characterization of, inter alia, tectal neurons in the monkey (Moschovakis et al. 1988), spinal motoneurons in the turtle (Hornby et al. 2002), and brainstem motoneurons in the frog (Matesz et al. 1995; Birinyi et al. 2004). Unfortunately, contrasting our present data with previous findings available in the literature could be performed to a limited extent only, because most of the parameters we determined have not been assessed before. Nevertheless, data characterizing the size of the cell bodies of pyramidal and giant cells could be compared to those published in several previous articles. When rhodamine backfilling was applied in the rat, the somatic diameter was 18–27 and 27–64 lm in the cases of pyramidal and giant cells, respectively (Pocsai et al. 2007). These values are similar to those obtained by Alibardi (1999)—24 lm for pyramidal and 28 lm for giant neurons. In cat, using Golgi impregnation, mature pyramidal cells had a size of 20 9 35 lm, whereas the maximum diameter of giant neurons was approximately 40 lm (Kane 1974). In Nissl-stained preparation, the diameter of the cat giant cells was defined as longer than 22 lm (Kane et al. 1981). The size of the giant cells was *30 lm in both gerbil (Ding et al. 1999) and mouse (Zhang and Oertel 1993). It can be concluded, therefore, that the soma diameters of pyramidal and giant cells may be similar, although the somewhat bigger maximum diameter of the giant cells is an unequivocal observation. With this regard, our present finding indicating that the longest diameter of the pyramidal neurons is significantly larger than that of the giant neurons is surprising. The most straightforward explanation of this discrepancy would be inadequate cell classification in the present work. However,
123
1290
we always performed the cell identification after confocal visualization and 3D reconstruction of the investigated neurons, and we always considered several parameters for the determination of the cell type, including the shape and localization of the cell body as well as the characteristic features of the entire dendritic arborization. Consequently, the unexpected finding about the size range of the giant and pyramidal cell bodies cannot be an artifact. This conclusion is supported by our observation that the surface area of the giant cell bodies was bigger than that of the pyramidal cells. Although the explanation of the aforementioned discrepancy is not clear, it should be pointed out that in one of our preceding morphological studies—due to technical limitations associated with rhodamine filling, such as inappropriate visualization of the dendritic tree—only cells possessing three dendritic trunks were considered as pyramidal neurons (Pocsai et al. 2007). This protocol most likely resulted in a somewhat biased cell selection, because only a subpopulation of the pyramidal cells was considered. Another explanation may be the different age of the rats employed in the two studies: in the present work 10–14 days old rats were used, whereas in our earlier article, more mature (1 month old) ones were studied (Pocsai et al. 2007). Although no data are available about the morphological consequences of the postnatal maturation of the CN in rats, detailed analyses of maturationassociated changes in hamster indicate that pyramidal cells considerably change their appearance during the first several weeks of the postnatal life (Schweitzer and Cant 1985; Schweitzer 1991). The shape of the cell bodies has been assessed quantitatively in the cat (Kane 1974; Kane et al. 1981). In these studies, the ratio of the shortest and longest diameters was used, which is the reciprocal of the ‘roundness’ we determined in the present work. After making the necessary adjustments, the quantitative data provided by the referenced articles correspond to a roundness of 1.5–1.7 and 1–1.5 for pyramidal and giant cells, respectively. These values are somewhat smaller than those reported in our present work, and especially the pyramidal cells seem to be more rotund in the cat. Nevertheless, the tendency established in the present analysis is rather similar to that reported in the cat, and it is in harmony with the generally accepted view that giant neurons have less elongated somata. There is a consensus view that giant neurons possess a larger number of initial dendritic processes than pyramidal cells do. In rhodamine-filled sections, positively identified giant cells had 3–5 stem dendrites (Pocsai et al. 2007), which is similar to that reported in the present study. As for the characteristics of the dendritic trees, there are only sporadic quantitative data for comparison. In Golgiimpregnated pyramidal cells of the cat, for example, the
123
Brain Struct Funct (2013) 218:1279–1292
diameter of the stem dendrites was 5–12 lm (Kane 1974), which is bigger than the same parameter obtained in the present analysis. The total dendritic length of the pyramidal cells of the cat was 6,536 lm (Blackstad et al. 1984), which is reasonably close to our finding (7,379 lm for completely reconstructed pyramidal cells; Table 2). Although we considered 17 parameters for the characterization of completely reconstructed neurons, only four of them showed statistically significant difference between giant and pyramidal cells. The ratio was somewhat better when only the somata and stem dendrites were considered, as in this case four of the eight parameters were statistically different. When comparing only the dendritic trees of giant and pyramidal cells, three parameters of the six exhibited statistically significant differences. These observations indicate that univariate quantitative analysis may not be sufficient for the adequate classification of the pyramidal and giant cells. In contrast, when discriminant analysis was employed—thus, all parameters were taken into account— the expected three distinct groups of the fully reconstructed neurons could be confirmed without identification mistakes. The multivariate analysis that was based on parameters describing the somata and the stem dendrites was reliable enough for distinguishing between the three cell groups even in the cases of incompletely reconstructed cells. We conclude that pyramidal and giant cells may be distinguished from other cell types of the DCN on the basis of their cell body diameter, while the soma roundness and the number of stem dendrites are the best discriminators when seeking the most reliable difference between giant and pyramidal neurons. Implications and potential use of the presented data Our results indicate that the combination of intracellular labeling, confocal microscopy, 3D reconstruction, and quantitative characterization of the neuronal morphology are helpful for the classification of neurones accommodated by the DCN. Some other implications of our present work are listed as follows. Firstly, the application of the multivariate analysis of several morphological parameters characterizing the soma and stem dendrites of pyramidal and giant cells may be particularly useful for cell identification after microelectrode experiments when the labeling does not visualize the entire dendritic tree. This is important because functional studies are often conducted on either pyramidal neurons only (e.g., Kanold and Manis 1999; Street and Manis 2007; Meng et al. 2012) or on ‘projection cells’ of the DCN in general (e.g., Ma and Brenowitz 2012). Secondly, Josephson and Morest (1998) have raised the intriguing possibility that individual neurons of the CN may not necessarily belong to discretely determined cell
Brain Struct Funct (2013) 218:1279–1292
populations but they may form a ‘‘continuum of morphological variability’’. According to this theory, intermediate cell forms may exist whose precise classification would be difficult—if at all possible. The similarity between the cell bodies, dendritic trees, and functional properties of the pyramidal and giant cells may be regarded as pieces of evidence supporting this ‘continuum’ hypothesis. However, our present results indicate that, if a large number of parameters are determined and appropriate combinations of these parameters are used to differentiate between cell types, it is possible to classify individual cells as belonging to one or the other group with reasonable confidence—at least in the cases of pyramidal and giant cells. Nevertheless, it will require further work to determine whether the ‘continuum’ hypothesis can be confirmed (or rejected) in other types of neurons. Thirdly, age-dependent changes of the functional and/or morphological properties of different types of neurons of the CN have already been described (e.g., Kane 1974; Schweitzer and Cant 1985; Schweitzer 1991; Cuttle et al. 2001; Caminos et al. 2005; Bortone et al. 2006). To have a better understanding of the nature and dynamics of these age-dependent (probably maturation-driven) changes, one must be aware of the cellular morphology and function at different time points during the postnatal life. Our present data obtained from 10- to 14-day-old rats may serve as reference in further studies exploring age-dependent changes of the principal cell types of the DCN. Finally, Tyner (1975) has suggested that the smallest units of brain function are ‘‘groups of similar neurons’’ rather than individual cells. It follows that the function of the CN is easier to comprehend if the observed functional properties are related to neuronal groups (defined on the basis of their common properties) as opposed to describing the features of sole, independent, or unclassified cells. For this reason, one of the primary aims of our present work was to provide a firm morphological basis for objective classification of giant and pyramidal cells. Acknowledgments This work was supported by grants from the Hungarian Scientific Research Fund (OTKA K-72812, OTKA K-67641), the Hungarian Academy of Sciences (MTA-TKI-355), and by an NHMRC (National Health & Medical Research Council) Australia Fellowship Grant awarded to George Paxinos (Grant #568605). Conflict of interest of interest.
The authors declare that they have no conflict
References Alibardi L (1999) Fine structure, synaptology and immunocytochemistry of large neurons in the rat dorsal cochlear nucleus connected to the inferior colliculus. J Hirnforsch 39:429–439
1291 Babalian AL, Ryugo DK, Rouiller EM (2003) Discharge properties of identified cochlear nucleus neurons and auditory nerve fibers in response to repetitive electrical stimulation of the auditory nerve. Exp Brain Res 153:452–460 Birinyi A, Szekely G, Csapo K, Matesz C (2004) Quantitative morphological analysis of the motoneurons innervating muscles involved in tongue movements of the frog Rana esculenta. J Comp Neurol 470:409–421 Blackstad TW, Osen KK, Mugnaini E (1984) Pyramidal neurones of the dorsal cochlear nucleus: a Golgi and computer reconstruction study in cat. Neuroscience 13:827–854 Bortone DS, Mitchell K, Manis PB (2006) Developmental time course of potassium channel expression in the rat cochlear nucleus. Hear Res 211:114–125 Caminos E, Vale C, Lujan R, Martinez-Galan JR, Juiz JM (2005) Developmental regulation and adult maintenance of potassium channel proteins (Kv 1.1 and Kv 1.2) in the cochlear nucleus of the rat. Brain Res 1056:118–131 Cant NB, Benson CG (2003) Parallel auditory pathways: projection patterns of the different neuronal populations in the dorsal and ventral cochlear nuclei. Brain Res Bull 60:457–474 Cant NB, Gaston KC (1982) Pathways connecting the right and left cochlear nuclei. J Comp Neurol 212:313–326 Cuttle MF, Ruszna´k Z, Wong AY, Owens S, Forsythe ID (2001) Modulation of a presynaptic hyperpolarization-activated cationic current (Ih) at an excitatory synaptic terminal in the rat auditory brainstem. J Physiol 534:733–744 Davis KA, Young ED (2000) Pharmacological evidence of inhibitory and disinhibitory neuronal circuits in dorsal cochlear nucleus. J Neurophysiol 83:926–940 Ding J, Benson TE, Voigt HF (1999) Acoustic and current-pulse responses of identified neurons in the dorsal cochlear nucleus of unanesthetized, decerebrate gerbils. J Neurophysiol 82: 3434–3457 Fredrich M, Reisch A, Illing RB (2009) Neuronal subtype identity in the rat auditory brainstem as defined by molecular profile and axonal projection. Exp Brain Res 195:241–260 Friauf E (1994) Distribution of calcium-binding protein calbindinD28k in the auditory system of adult and developing rats. J Comp Neurol 349:193–211 Frisina RD, Zettel ML, Kelley PE, Walton JP (1995) Distribution of calbindin D-28k immunoreactivity in the cochlear nucleus of the young adult chinchilla. Hear Res 85:53–68 Hancock KE, Voigt HF (2002) Intracellularly labeled fusiform cells in dorsal cochlear nucleus of the gerbil. II. Comparison of physiology and anatomy. J Neurophysiol 87:2520–2530 Heiman-Patterson TD, Strominger NL (1985) Morphological changes in the cochlear nuclear complex in primate phylogeny and development. J Morphology 186:289–306 Hornby TG, McDonagh JC, Reinking RM, Stuart DG (2002) Electrophysiological properties of spinal motoneurons in the adult turtle. J Comp Physiol 188:97–408 Josephson EM, Morest DK (1998) A quantitative profile of the synapses on the stellate cell body and axon in the cochlear nucleus of the chinchilla. J Neurocytol 27:841–864 Juiz JM, Luja´n R, Domı´nguez del Toro E, Fuentes V, Ballesta JJ, Criado M (2000) Subcellular compartmentalization of a potassium channel (Kv1.4): preferential distribution in dendrites and dendritic spines of neurons in the dorsal cochlear nucleus. Eur J Neurosci 12:4345–4356 Kane EC (1974) Synaptic organisation in the dorsal cochlear nucleus of the cat: a light and electron microscopic study. J Comp Neurol 155:301–330 Kane ES, Puglisi SG, Gordon BS (1981) Neuronal types in the deep dorsal cochlear nucleus of the cat: I. Giant neurons. J Comp Neurol 198:483–513
123
1292 Kanold PO, Manis PB (1999) Transient potassium currents regulate the discharge patterns of dorsal cochlear pyramidal cells. J Neurosci 19:2195–2208 Korada S, Schwartz IR (2000) Calcium binding proteins and the AMPA glutamate receptor subunits in gerbil cochlear nucleus. Hear Res 140:23–37 Kudo M, Nakamura K (1987) Organization of the lateral lemniscal fibers converging onto the inferior colliculus in the cat: an anatomical review. In: Syka J, Masterton RB (eds) Auditory pathway: structure and function. Plenum Press, New York, pp 171–183 Lorente de No R (1981) The primary acoustic nuclei. Raven Press, New York Ma WLD, Brenowitz SD (2012) Single-neuron recordings from unanesthetized mouse dorsal cochlear nucleus. J Neurophysiol 107:824–835 Malmierca MS (2003) The structure and physiology of the rat auditory system: an overview. Int Rev Neurobiol 56:147–211 Manis PB (1990) Membrane properties and discharge characteristics of guinea pig dorsal cochlear nucleus neurons studied in vitro. J Neurosci 10:2338–2351 Matesz C, Birinyi A, Kothalawala DS, Szekely G (1995) Investigation of the dendritic geometry of brain stem motoneurons with different functions using multivariant statistical techniques in the frog. Neuroscience 65:1129–1144 Meng X, Lu Q, Rinzel J (2012) Control of firing patterns by two transient potassium currents: leading spike, latency, bistability. J Comput Neurosci 31:117–136 Moschovakis AK, Karabelas AB, Highstein SM (1988) Structure– function relationships in the primate superior colliculus. I. Morphological classification of efferent neurons. J Neurophysiol 60:232–262 Oliver DL (1984) Dorsal cochlear nucleus projections to the inferior colliculus in the cat: a light and electron microscopic study. J Comp Neurol 224:155–172 Osen KK (1969) Cytoarchitecture of the cochlear nuclei in the cat. J Comp Neurol 136:453–484 Osen KK (1972) Projection of the cochlear nuclei on the inferior colliculus in the cat. J Comp Neurol 144:355–372 Ostapoff EM, Feng JJ, Morest DK (1994) A physiological and structural study of neuron types in the cochlear nucleus. II. Neuron types and their structural correlation with response properties. J Comp Neurol 346:19–42 ´ , Szu¨cs G, Kova´cs I, Ruszna´k Z (2003) HCN channels Pa´l B, Po´r A contribute to the intrinsic activity of cochlear pyramidal cells. Cell Mol Life Sci 60:2189–2199 ´ , Pap P, Bakondi G, Pocsai K, Szu¨cs G, Ruszna´k Z Pa´l B, K}oszeghy A (2009) Targets, receptors and effects of muscarinic neuromodulation on giant neurones of the rat dorsal cochlear nucleus. Eur J Neurosci 30:769–782 Perney TM, Kaczmarek LK (1997) Localization of a high threshold potassium channel in the rat cochlear nucleus. J Comp Neurol 386:178–202 Pocsai K, Pa´l B, Pap P, Bakondi G, Kosztka L, Ruszna´k Z, Szu¨cs G (2007) Rhodamine backfilling and confocal microscopy as a tool for the unambiguous identification of neuronal cell types: a study of the neurones of the rat cochlear nucleus. Brain Res Bull 71:529–538 ´ , Pocsai K, Ruszna´k Z, Szu¨cs G (2005) Presence and distribution Po´r A of three calcium binding proteins in projection neurons of the adult rat cochlear nucleus. Brain Res 1039:63–74 Rhode WS, Smith PH, Oertel D (1983) Physiological response properties of cells labeled intracellularly with horseradish peroxidase in cat dorsal cochlear nucleus. J Comp Neurol 213:426–447 Rogers JH, Resibois A (1992) Calretinin and calbindin-D28k in rat brain: patterns of partial co-localization. Neuroscience 51:843–865
123
Brain Struct Funct (2013) 218:1279–1292 Rowe MH, Stone J (1977) Naming of neurons. Classification and naming of cat retinal ganglion cells. Brain Behav Evol 14:185–216 Ruszna´k Z, Forsythe ID, Brew HM, Stanfield PR (1997) Membrane currents influencing action potential latency in granule neurons of the rat cochlear nucleus. Eur J Neurosci 9:2348–2358 Ruszna´k Z, Bakondi G, Pocsai K, Por A, Kosztka L, Pal B, Nagy D, Szu¨cs G (2008) Voltage-gated potassium channel (Kv) subunits expressed in the rat cochlear nucleus. J Histochem Cytochem 56:443–465 Ryugo DK, Willard FH (1985) The dorsal cochlear nucleus of the mouse: a light microscopic analysis of neurons that project to the inferior colliculus. J Comp Neurol 242:381–396 Ryugo DK, Willard FH, Fekete DM (1981) Differential afferent projections to the inferior colliculus from the cochlear nucleus in the albino mouse. Brain Res 210:342–349 Schofield BR, Cant NB (1996) Origins and targets of commissural connections between the cochlear nuclei in guinea pigs. J Comp Neurol 375:128–146 Schofield BR, Coomes DL (2005) Projections from auditory cortex contact cells in the cochlear nucleus that project to the inferior colliculus. Hear Res 206:3–11 Schweitzer L (1990) Differentiation of apical, basal and mixed dendrites of fusiform cells in the cochlear nucleus. Brain Res Dev Brain Res 56:19–27 Schweitzer L (1991) Morphometric analysis of developing neuronal geometry in the dorsal cochlear nucleus of the hamster. Brain Res Dev Brain Res 59:39–47 Schweitzer L, Cant NB (1985) Differentiation of the giant and fusiform cells in the dorsal cochlear nucleus of the hamster. Brain Res 352:69–82 Sholl DA (1953) Dendritic organization in the neurons of the visual and motor cortices of the cat. J Anat 87:387–406 Shore SF, Godfrey DA, Helfert RH, Altschuler RA, Bledsoe SC Jr (1992) Connections between the cochlear nuclei in guinea pig. Hear Res 62:16–26 Smith PH, Massie A, Joris PX (2005) Acoustic stria: anatomy of physiologically characterized cells and their axonal projection patterns. J Comp Neurol 482:349–371 Spatz WB (1997) Differences between guinea pig and rat in the dorsal cochlear nucleus: expression of calcium-binding proteins by cartwheel and Purkinje-like cells. Hear Res 107:136–146 Street SE, Manis PB (2007) Action potential timing precision in dorsal cochlear nucleus pyramidal cells. J Neurophysiol 97:4162–4172 Tyner CF (1975) The naming of neurons: applications of taxonomic theory to the study of cellular populations. Brain Behav Evol 12:75–96 Wearne SL, Rodriguez A, Ehlenberger DB, Rocher AB, Henderson SC, Hof PR (2005) New techniques for imaging, digitization and analysis of three-dimensional neural morphology on multiple scales. Neuroscience 136:661–680 Webster DB, Trune DR (1982) Cochlear nuclear complex of mice. Am J Anat 163:103–130 Wright DD, Blackstone CD, Huganir RL, Ryugo DK (1996) Immunocytochemical localization of the mGluR1 alpha metabotropic glutamate receptor in the dorsal cochlear nucleus. J Comp Neurol 364:729–745 Young ED, Spirou GA, Rice JJ, Voigt HF (1992) Neural organization and responses to complex stimuli in the dorsal cochlear nucleus. Phil Trans R Soc Lond B336:407–413 Zhang S, Oertel D (1993) Giant cells of the dorsal cochlear nucleus of mice: intracellular recordings in slices. J Neurophysiol 69:1398–1408 Zhang S, Oertel D (1994) Neuronal circuits associated with the output of the dorsal cochlear nucleus through fusiform cells. J Neurophysiol 71:914–930 Zheng X, Voigt HF (2006) Computational model of response maps in the dorsal cochlear nucleus. Biol Cybern 95:233–242