* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Refinement of feedforward projections, neuronal density, and
Holonomic brain theory wikipedia , lookup
Human brain wikipedia , lookup
Types of artificial neural networks wikipedia , lookup
Cortical cooling wikipedia , lookup
Nonsynaptic plasticity wikipedia , lookup
Haemodynamic response wikipedia , lookup
Neuropsychopharmacology wikipedia , lookup
Subventricular zone wikipedia , lookup
Axon guidance wikipedia , lookup
Neuroplasticity wikipedia , lookup
Neuroeconomics wikipedia , lookup
Apical dendrite wikipedia , lookup
Neuroesthetics wikipedia , lookup
Convolutional neural network wikipedia , lookup
Environmental enrichment wikipedia , lookup
Activity-dependent plasticity wikipedia , lookup
Premovement neuronal activity wikipedia , lookup
Neuroanatomy wikipedia , lookup
Optogenetics wikipedia , lookup
Anatomy of the cerebellum wikipedia , lookup
Development of the nervous system wikipedia , lookup
Nervous system network models wikipedia , lookup
Neural correlates of consciousness wikipedia , lookup
Channelrhodopsin wikipedia , lookup
Metastability in the brain wikipedia , lookup
Synaptic gating wikipedia , lookup
Aging brain wikipedia , lookup
Feature detection (nervous system) wikipedia , lookup
Cerebral cortex wikipedia , lookup
City University of New York (CUNY) CUNY Academic Works Master's Theses City College of New York 2015 Refinement of feedforward projections, neuronal density, and characterization of synapsesin layer 4 of ferret primary visual cortex Violeta Contreras Ramirez CUNY City College How does access to this work benefit you? Let us know! Follow this and additional works at: http://academicworks.cuny.edu/cc_etds_theses Part of the Biology Commons Recommended Citation Ramirez, Violeta Contreras, "Refinement of feedforward projections, neuronal density, and characterization of synapsesin layer 4 of ferret primary visual cortex" (2015). CUNY Academic Works. http://academicworks.cuny.edu/cc_etds_theses/357 This Thesis is brought to you for free and open access by the City College of New York at CUNY Academic Works. It has been accepted for inclusion in Master's Theses by an authorized administrator of CUNY Academic Works. For more information, please contact [email protected]. Refinement of feedforward projections, neuronal density, and characterization of synapses in layer 4 of ferret primary visual cortex Violeta Contreras Ramirez Mentor: Dr. Jonathan B. Levitt A thesis submitted in partial fulfillment of the requirements for the BS/MS dual degree program in the Biology Department of the City College of New York of the City University of New York 2015 Introduction We are interested in studying the organization and refinement of brain circuits in the visual cortex. To accomplish this, we used a combination of light and electron microscope (EM) techniques to quantify how various connectional characteristics in visual areas change throughout development in the ferret. Having a richer understanding of the maturation of different visual areas is relevant to understanding the neural correlates of the maturation of different visual functions, since these do not all develop at the same age. For example, in humans, cells show responses that are orientation-selective at around 6 weeks of age, but responses to direction of motion do not occur until 10-12 weeks of age (Braddick et al., 1986; Wattam-Bell, 1991). The developing visual cortex is a very interesting model to study due to its ability to be greatly affected by experience. The importance of experience in the refinement of visual connections is illustrated by the finding that rats placed in a complex environment display 27% more synapses per neuron than rats in a simple environment (Briones et al., 2004). Furthermore, kittens with normal vision have a large portion of their visual cortex that is binocular, and thus, does not show bias in eye specificity. Causing strabismus in kittens for 2 days leads to a smaller binocular zone, while strabismus lasting for 7-14 days completely eliminates it, resulting in a cortex that is biased for same-eye domains (Trachtenberg and Stryker, 2001). Finally, it has been found that preventing the contralateral eye from opening at P9 in mice retards maturation of ipsilateral retinotopy. In contrast, enucleating the contralateral eye leads to acceleration in retinotopic map maturation by P15. Vision is necessary for this acceleration, as enucleating the contralateral eye and preventing the ipsilateral eye from opening until P15 results in a retinotopic 1 map that is less organized than that found when the contralateral eye is enucleated but the ipsilateral eye is allowed to open naturally between P10 to P12 (Smith and Trachtenberg, 2007). While there are features whose development is clearly dependent on visual experience, other features are partially intrinsic; but even in these cases, visual experience might still play a role. Ko et al. (2014) found that mice that are deprived of vision from before eye opening still develop orientation-selective cells in V1, just as controls do. In addition, there is an increased probability for cells in layers 2 and 3 to form connections with other cells with similar orientations, compared to with cells responding to orthogonal orientations, and this is seen in dark-reared mice as well as controls. Nevertheless, visually unresponsive cells lose connections made to visually responding cells in controls, but this loss is not present in dark reared mice (Ko et al., 2014). Furthermore, the orientation selectivity present in visually deprived mice could have developed as a result of visual experience before the mice opened their eyes. Krug et al. (2001) provide evidence for orientation-selectivity in LGN and V1 neurons in ferrets in the two weeks before eye opening. Their data suggest that visual stimuli could be involved in the development of orientation selectivity even through closed eyelids. Thus, visual experience could still be shaping the development of features that initially appear to be intrinsic. A possible mechanism for the role of visual experience in refining connections in the visual cortex may involve the action of astrocytes. Astrocytes are known to be involved in guiding axons to their destinations, as well as in synapse formation and elimination. Since neurogenesis occurs earlier than gliogenesis in the developing cortex, it is believed that neurons may provide extrinsic cues that allow neural stem cells to differentiate into glial cells. In particular, the IL-6 family of cytokines may promote the formation of astrocytes (Sloan and Barres, 2014). Thus, it is reasonable to expect that cortical neurons receiving visual stimuli 2 would be activated more than non-visually activated neurons, and the former, more active neurons will produce more cytokines, leading to the proliferation of more astrocytes. Astrocytes, in turn, will be important for the elimination of synapses. Thus, this mechanism could explain how visual experience leads to refinement of connections in the visual cortex. A fundamental question in neuroscience is whether different visual areas mature with similar or different timecourses. In order to shed light on this question, we quantified changes in the density and distribution of boutons in feedforward projections, which are those that originate in area 17 and extend to extrastriate areas (Callaway, 2004). By studying the distribution of synaptic zinc in the ferret visual cortex, Khalil and Levitt (2013) found evidence for a similar developmental timecourse of different visual areas. Synaptic zinc is present in a subset of glutamatergic processes from various sources, and the intensity of the label declines with age. The major decline occurred from 5 to 6 weeks, which is around the time of eye opening in the ferret (P32). However, we are now interested in studying a specific type of glutamatergic processes: feedforward projections. To do this, we used a tracer to label axons that project from area 17 to extrastriate areas. We assessed the developmental timecourse of extrastriate areas by quantifying bouton density, interbouton intervals, and axon density. Boutons are the anatomical specializations by which neurons make synaptic contacts, and interbouton intervals refer to the distance between adjacent boutons along individual axons. Thus, by quantifying boutons, we can obtain an estimate of how connectivity changes during development. In addition, the interbouton intervals and axon density allow us to determine if changes in bouton density are due to the elimination of individual boutons, the pruning of axonal branches, or both of these processes. Given that connections can be eliminated while the neuron survives, we decided to compare the developmental changes in bouton density to those of neuronal density in different 3 visual areas. It is well established that the brain contains exuberant connections early in development, and thus, we expected to find a decrease in feedforward projections with age. Exuberance refers to the overproduction of axons and synapses and their selection for persistence during development. Elimination of projections may result from neuronal death or the elimination of synapses, and selection is determined by input from the periphery, competition among axons, and hormones. It is believed that exuberance is important for modulating the strength of connections and providing circuits with flexibility. Exuberance occurs in both cortico-cortical and cortico-subcortical projections (Innocenti and Price, 2005). To compare changes in bouton density with those in neuronal density, we obtained the latter measurement at 4 and 6 weeks of age, and adulthood. While the pruning of exuberant projections leads to a decrease in connections during development, we expected to also find a decrease in neuronal density due to the enlargement of the brain with age (and not due to cell death). The weight of the ferret brain, which is indicative of its size, increases at a very fast rate from P0 to P30. However, even after this age, it continues to increase in weight. From 4 weeks to 6 weeks, the weight increases from approximately 5 to 7 grams, an increase of 40% (Barnette et al., 2009). This increase in brain weight, and size, is coupled with a lack of neurogenesis. It has been reported that in ferrets, neurogenesis in the visual cortex ends at approximately 2 weeks after birth (Jackson et al., 1989). Thus, if the brain is getting bigger, and no more neurons are being produced, then neuronal density should decline. Having quantified a proxy for synapses, we then asked what particular aspects of synapses are changing with age; thus, a final analysis we undertook was to use electron microscopy to characterize different types of synapses in layer 4 of area 17 in the ferret. While we hope to quantify synaptic changes during development at the EM level, we began by 4 describing several synaptic features in the adult, in order to understand what the organization of the developing cortex will ultimately be. In addition to synapse density, we identified the postsynaptic targets as dendritic shafts or spines. While synapses can occur on both of these structures, spines are the most common targets (DeFelipe et al., 1999). We also differentiated between asymmetric and symmetric synapses. This categorization is important, as asymmetric synapses are thought to be excitatory, and symmetric are thought to be inhibitory, with the former being more common (DeFelipe et al., 1999). Thus, by quantifying these anatomical characteristics, we can gain insight into the functional circuits in primary visual cortex. After quantifying the above-mentioned characteristics in all types of synapses of layer 4 in V1, we differentiated between the measurements of cortico-cortical and thalamo-cortical synapses. Given that the synapses we are interested in classifying are glutamatergic, they use a glutamate transporter to recycle the neurotransmitter, and different populations of cells use different glutamate transporters. The vglut1 transporter is believed to be used in cortico-cortical synapses, while vglut2 is used in thalamocortical synapses. Indeed, Nahmani and Erisir (2005) found that vglut2 immunoreactivity (ir) can effectively identify thalamo-cortical projections to layer 4 of V1. Furthermore, they found that vglut2 is not used by cortico-cortical synapses, and that vglut2 is present in boutons that form asymmetric synapses, indicating that these are glutamatergic. Finally, these researchers identified that vglut2+ boutons have an even greater tendency than cortico-cortical boutons to synapse onto spines compared to dendritic shafts (77% vs. 59%) (Nahmani and Erisir, 2005). While vglut2+ puncta can be found in any layer of V1 of monkey to different degrees, they are much more heavily concentrated in layer 4Cβ (GarciaMarin et al., 2012). Therefore, we were able to use vglut2 immunoreactivity to identify thalamo- 5 cortical synapses in layer 4 of V1 and compare the characteristics of these synapses with those of cortico-cortical synapses in the same layer. The reason for differentiating between particular types of synapses in specific layers is that the visual cortex is very highly organized anatomically, and as such, a “canonical circuit” has been proposed to describe the functional properties of the cortex. A canonical circuit is a microcircuit repeated in all areas of the cortex, where it performs similar functional computations (Douglas et al., 1989). To understand the significance of the canonical circuit, it is useful to review the basic organization of the visual cortex, as this was the first area used to describe the circuit. Very broadly, neurons in the cortex can be divided into excitatory and inhibitory classes, with the former representing the majority of neurons in V1 (Binzegger et al., 2004). While the majority of excitatory neurons are pyramidal and glutamatergic, inhibitory neurons have smooth dendrites (do not have spines) and are GABAergic. Even though the much larger proportion of excitatory to inhibitory cells holds across several species (cat, monkey, and rodents), in rodents, pyramidal cells in layers 2 and 3 target more inhibitory neurons than we would expect by chance (Bopp et al., 2014). Nevertheless, it is generally the case that most connections in V1 are from layer 2/3 pyramidal cells onto other pyramidal cells in the same layers (Binzegger et al., 2009). Layer 1 consists mainly of axons, dendrites, and glial processes, while layers 2, 3, 5, and 6 are predominantly made up of pyramidal neurons. Layer 4, on the other hand, has spiny stellate neurons in the largest proportion (DeFelipe et al., 1999). Nevertheless, this layer also contains GABAergic neurons, of which 78% are basket cells (Binzegger et al., 2004). Layer 4 is unique in that it is the primary input layer for thalamo-cortical synapses from the lateral geniculate nucleus (LGN). Despite this, layer 4 neurons receive many more cortico-cortical synapses than thalamocortical synapses. Indeed, layer 6 pyramidal neurons and layer 4 spiny stellate neurons provide 6 approximately 8 and 5 times as many synapses onto layer 4 neurons as LGN neurons do (Ahmed et al., 1994). Despite its proportionately small input, it is thought that LGN input to V1 accounts for most of the orientation tuning of V1 cells (Lamme et al., 1998). Indeed, the model of the canonical circuit suggests that thalamocortical inputs into layer 4 neurons drive these cells, whose responses then get amplified by cortico-cortical excitation. GABAergic inhibition from cortical cells onto excitatory cells in all layers of the cortex is provided to prevent overexcitation (Douglas et al., 1989). Even though they were first described in the visual cortex, canonical circuits perform similar functional computations throughout the cortex, as the specific wiring of the canonical circuit is flexible and can be universally applied (Costa and Martin, 2010). Indeed, differences in the laminar organization of different areas require a labile definition of a functional unit, as Beul and Hilgetag (2015) argue in trying to build a canonical circuit that is consistent with less organized regions than V1, such as the prefrontal cortex, which contains agranular regions. While the canonical circuit model emphasizes the importance of thalamo-cortical and intra-areal connections, it is important to keep in mind that another set of important inputs to V1 is that from extrastriate areas: feedback projections. It has been shown that both feedforward and feedback projections are mostly excitatory (Lamme et al., 1998). While thalamo-cortical projections to V1 drive these neurons, feedback projections have been suggested to modulate V1 neurons’ responses (Angelucci et al., 2002; Budd, 1998). For example, Hupe et al. (1998) found that inactivating area V5 by cooling leads to a decrease in the response of V1, V2, and V3 neurons. Their data also suggest that V5 is involved in enhancing responses to the figure relative to the background in lower visual areas. V2, or area 18, is the largest extrastriate area and provides more input to V1 than any other extrastriate area. Nevertheless, this feedback input 7 accounts for only approximately 0.9-6.1% of excitatory inputs onto V1 neurons in layers 2 and 3 (Budd, 1998). Our study focused on the refinement and organization of the visual cortex of the ferret. We calculated neuronal density, bouton density, interbouton intervals, and axon density of feedforward terminal clusters at different developmental stages, and we characterized synapses in layer 4 of area 17 in adults. We chose the ferret (Mustela putorius furo) for this study because its brain is very underdeveloped at birth (Sur and Leamy, 2001). In fact, the emergence of layers 2/3 and 4, as well as connections from the LGN to the cortex occur postnatally in ferrets. The ages at which we studied the refinement of the visual cortex in ferrets were chosen to span the period around eye opening, which is approximately P30 to P32 (Durack and Katz, 1996). Thus, we could observe if visual experience could be driving the development of connections among visual areas, and whether different areas refine synchronously or asynchronously. We hypothesized that different visual areas would develop with different timecourses, as this could indicate the neural correlates of the different maturation rates of distinct visual functions. Interestingly, it is during and shortly after the period of eye opening during which we found the greatest refinement. However, contrary to our hypothesis, we found that many aspects of refinement that we observed occurred at similar rates in different visual areas. Methods General methods We studied the visual cortex of ferrets at different developmental stages. For neuronal density, peak bouton density, and interbouton intervals, we used a sample size of two ferrets per area at most ages. We used a sample size of one for peak bouton density and intervals at 5 8 weeks; so far, we have studied only one ferret at each age for the calculations of overall bouton density and axon density. The ferrets were obtained from Marshal Farms (North Rose, NY) and maintained under a 12 hour light/dark cycle. Anatomical tracer injections We injected the tracer cholera toxin B (CTb) into area 17 (following the protocol described in Khalil and Levitt, 2014), and after a week of survival, the animals were euthanized with an intraperitoneal administration of pentobarbital (60-100 mg/kg). They were then perfused with saline and their brains were fixed with a 4% paraformaldehyde solution. The ferrets’ brains were removed and frozen sections were tangentially cut into 40 µm thick sections using a sliding microtome. The sections were mounted on slides, dehydrated, cleared in xylene, and coverslipped. While alternate sections were reacted to reveal the CTb label, other sections were processed for Nissl substance, synaptic zinc, or cytochrome oxidase (Cantone et al., 2005). Bouton and axon density measurements We used a light microscope and the program Neurolucida (MBF Bioscience: Williston VT) to mark boutons and axons in the ferret extrastriate cortex and calculate peak bouton density, overall bouton density, interbouton intervals, and axon density. We found peak and overall bouton density, interbouton intervals, and axon density in areas 18, 19, 21, and Ssy at 4, 5, 6, 8, and 10 weeks of age (we performed measurements in a 7 week old case as well, but we excluded the data due to sparse labeling in the animal). To calculate peak bouton density, we counted the number of boutons in 5 μm by 5 μm boxes using a 100x oil immersion lens to view the region of interest. We placed a total of 10 of these sample boxes in the core, mid-periphery, and periphery of each CTb labeled cluster in each area at each age in each animal. We counted 9 the boutons along the depth of each section using a 1μm exclusion zone at the top and bottom of the Z plane. In addition, we used the bottom and left edges of each sample box as exclusion zones to prevent overcounting. We followed a similar procedure for estimating overall bouton density, with the exception that we specifically looked for label when placing our sample boxes in the peak bouton density measurements, but not when estimating overall bouton density. Instead, for overall bouton density, we laid a grid of 5 μm by 5 μm boxes over our area of interest, and we used random.org to generate a random number from 1-14, which indicated the first box from which we took measurements (we numbered boxes from left to right). Subsequently, we sampled from every 14th box and counted from as many boxes as needed in order to obtain at least 100 boutons per region, per area, in each case. To calculate axon density, we measured the length of all axon processes present in 10 μm by 10 μm boxes at 40x. As in the peak bouton density counts, we placed 10 boxes in each region (core, mid-periphery, and periphery), in each area, at each age. When calculating densities, we corrected for shrinkage, as the tissue shrank due to processing by an average of 62%. We also used a 100x oil immersion lens to view the area of interest when we calculated interbouton intervals in the core and periphery of each CTb labeled cluster. We marked the boutons along individual axons, excluding terminal boutons to prevent overcounting, and used Neurolucida to find the distance between adjacent boutons in a given axon. Several axons were studied in each cluster, until we had accumulated at least 100 interbouton intervals in the core and 100 in the periphery in each area, at each age, in each case. Neuron counting 10 We calculated neuronal density in areas 17, 18, 19, 21 and Ssy using a light microscope at 100x oil immersion lens to view the tissue. We used Neurolucida to draw columns perpendicular to the cortical surface. These columns spanned layers 2 to 6, since layer 1 was excluded due to its low number of neurons. Each column contained a 45 µm by 45 µm square in each of the five layers studied. Neurons were identified by their distinct nucleoli, and counted along the X,Y, and Z planes of each box. In the Z plane, we used an exclusion zone of 2 µm at the top and the bottom to prevent over counting. In the X and Y planes of each box, the bottom and left lines were used as exclusion zones. The average tissue shrinkage in the Z plane was around 50%. Once the number of neurons in a box was determined, we calculated the density by dividing this number by the volume of the box in mm3 and correcting for shrinkage. Electron microscopy To have a richer understanding of the particular types of synapses present in the visual cortex, we characterized cortico-cortical and thalamo-cortical synapses in layer 4 of primary visual cortex. We used a total of 4 adult ferrets to categorize synapses at the EM level. The brains of 2 adult ferrets were processed for EM following the protocol in Garcia-Marin et al. (2009). The tissue from the other 2 brains was reacted with an antibody against the vesicular glutamate transporter (vglut2), following the protocol in Garcia-Marin et al. (2012), in order to allow us to distinguish cortico-cortical synapses from thalamo-cortical synapses, as the latter use the vglut2 transporter. The processing of these 4 adult brains was done by our collaborator Virginia Garcia-Marin at New York University. Using the program Fiji, we counted the number of synapses present in layer 4 of area 17 in randomly chosen EM images from all 4 animals, and characterized the synapses as asymmetric when there was a prominent PSD, as symmetric when the PSD was very thin, or as uncharacterized when it was unclear whether the synapse was 11 symmetric or asymmetric, when there was no clear synaptic cleft, or when the synapse was en face. We also measured the length of the PSD and identified the post-synaptic target as either a spine or a dendritic shaft. Results Developmental changes in bouton density We measured the peak bouton density in different visual areas at different ages to understand how the strength of connections in each area changes with age, and to elucidate if different areas might develop with similar or different timecourses. We obtained peak bouton densities in different regions of terminal clusters, namely: the core, mid-periphery, and periphery, as the boutons may not be distributed equally throughout the cluster. Figures 1a to 1c display the average bouton density ± S.E.M. in these different regions of the cluster in areas 18, 19, 21, and Ssy at 4, 5, 6, 8, and 10 weeks of age. In general, the core of each cluster contains the highest bouton density, particularly in the younger animals. In all parts of the cluster, there is more variability in the densities among areas early in development, while the densities at 8 and 10 weeks, with the exception of area 19 at the latter age, are nearly indistinguishable. In general, the bouton density in the different areas decreases with age, although the data also show increases from 5 to 6 weeks of age. Comparing the densities at 4 weeks with those at 10 weeks shows that there is a decline in most of the areas by at most 50%. We performed a permutation test and found significant differences at the P<0.05 level in bouton density with age in all areas except for area 19. Interestingly, all of the areas seem to follow the same trend, with peak densities at 4 and 6 weeks of age, and lower densities at all other ages. 12 Bouton Density (boutons x10^6/mm^3) Peak Bouton Density in the Core 25 20 15 18 19 10 21 Ssy 5 0 2 4 6 8 Age (weeks) 10 12 Figure 1a. Average peak bouton density in the core of the cluster in areas 18, 19, 21, and Ssy at different ages. Error bars represent S.E.M. Bouton Density (boutons x10^6/mm^3) Peak Bouton Density in the Mid-Periphery 25 20 15 18 19 10 21 5 Ssy 0 2 4 6 8 Age (weeks) 10 12 Figure 1b. Average peak bouton density in the mid-periphery of the cluster in areas 18, 19, 21, and Ssy at different ages. Error bars represent S.E.M. 13 Bouton Density (boutons x10^6/mm^3) Peak Bouton Density in the Periphery 25 20 15 18 19 10 21 Ssy 5 0 2 4 6 8 Age (weeks) 10 12 Figure 1c. Average peak bouton density in the periphery of the cluster in areas 18, 19, 21, and Ssy at different ages. Error bars represent S.E.M. Figures 2a to 2d display the same peak bouton density data as that plotted above, but these graphs allow us to more clearly observe the developmental changes occurring in each area. Figure 2a shows that in area 18, the core contains the highest density at 4, 6, and 8 weeks of age, while the densities are more similar in different parts of the cluster in the rest of the ages. In contrast, in area 19, the core contains the highest density and the mid-periphery the second highest at all ages. Area 21 shows a similar trend to that of area 19, with the exception that the 10 week olds have nearly indistinguishable densities in different regions of the cluster. As shown in 2d, the core and mid-periphery of the Ssy cluster contain very similar bouton densities from 4 to 6 weeks of age, while that of the periphery is substantially lower. Nevertheless, starting at 8 weeks of age, there is more overlap between the densities in all three regions of the cluster. All the areas show a similar trend of changes in density with age, with a relatively low density at 5 weeks, and a decline after 6 weeks of age. 14 Bouton Density (boutonsx10^6/mm^3) Peak Bouton Density in Area 18 25 20 15 core 10 mid-periphery periphery 5 0 2 4 6 8 Age (weeks) 10 12 Figure 2a. Average peak bouton density in the core, mid-periphery, and periphery of the cluster in area 18 as a function of age. Error bars represent S.E.M. Bouton Density (boutonsx10^6/mm^3) Peak Bouton Density in Area 19 25 20 15 core 10 mid-periphery periphery 5 0 2 4 6 8 Age (weeks) 10 12 Figure 2b. Average peak bouton density in the core, mid-periphery, and periphery of the cluster in area 19 as a function of age. Error bars represent S.E.M. 15 Bouton Density (bouonx10^6/mm^3) Peak Bouton Density in Area 21 25 20 15 core 10 mid-periphery periphery 5 0 0 2 4 6 8 Age (weeks) 10 12 Figure 2c. Average peak bouton density in the core, mid-periphery, and periphery of the cluster in area 21 as a function of age. Error bars represent S.E.M. Bouton Density (boutonsx10^6/mm^3) Peak Bouton Density in Ssy 25 20 15 core 10 mid-periphery periphery 5 0 2 4 6 8 Age (weeks) 10 12 Figure 2d. Average peak bouton density in the core, mid-periphery, and periphery of the cluster in Ssy as a function of age. Error bars represent S.E.M. Given that we obtained peak bouton density by sampling specifically from locations in the cluster that contained labeled boutons, we wondered if perhaps our methodology could have 16 masked certain developmental changes in bouton density occurring with age. Thus, we decided to take a completely random measurement of bouton density from the same three regions of each cluster of areas 18, 19, 21, and Ssy at 4, 5, 6, 8, and 10 weeks of age. Figures 3a to 3c show these measurements by region of the cluster as a function of age. While the values for overall bouton density are lower than those of peak bouton density, the trends are the same. A permutation test revealed changes in average bouton density as a function of age to be highly significant in all areas (P≤0.0006). Early in development, there is more variability among densities in different areas in all parts of the cluster. However, by 10 weeks of age, with the exception of the density in the mid-periphery of Ssy, the densities of all areas are essentially the same within a given region of the cluster. While the core and mid-periphery have very similar densities in all areas at all ages, the densities in the periphery are somewhat lower. However, this lower density in the periphery is present at 4 and 6 weeks of age, but not in the older animals. Bouton Density (10^6/mm^3) Overall Bouton Density in the Core 16 12 18 8 19 21 4 Ssy 0 4 6 8 10 Age (weeks) Figure 3a. Overall bouton density in the core of the cluster in areas 18, 19, 21, and Ssy at different ages. Error bars represent S.E.M. 17 Bouton Density (10^6/mm^3) Overall Bouton Density in the Mid-periphery 16 12 18 8 19 21 4 Ssy 0 4 6 8 10 Age (weeks) Figure 3b. Overall bouton density in the mid-periphery of the cluster in areas 18, 19, 21, and Ssy at different ages. Error bars represent S.E.M. Overall Bouton Density in the Periphery Bouton Density (10^6/mm^3) 16 12 18 8 19 21 4 Ssy 0 4 6 8 10 Age (weeks) Figure 3c. Overall bouton density in the periphery of the cluster in areas 18, 19, 21, and Ssy at different ages. Error bars represent S.E.M. Figures 4a through 4d display the same overall bouton density described above, but listed by area. All of the areas show similar trends in that the major bouton density declines occurs 18 after 6 weeks of age. In areas 18 and 21, however, there is a substantial decline in overall bouton density from 4 to 6 weeks of age in the core and mid-periphery of the cluster, but not in the periphery. In general, most regions of the cluster show a higher density at 10 weeks compared to 8 weeks in all of the areas. The only steady decline of bouton density from 4 to 10 weeks occurs in the periphery of Ssy, as all other regions in all areas show either an increase or little change from 8 to 10 weeks. Despite the small variation among regions of different areas at different ages, all of the areas show a bouton density decline of approximately 50% at 10 weeks compared to 4 weeks, and this is true in the core, mid-periphery, and periphery of each cluster. Bouton Density (10^6/mm^3) Overall Bouton Density in Area 18 16 12 core 8 mid-periphery 4 periphery 0 2 4 6 8 Age (weeks) 10 12 Figure 4a. Overall bouton density in the core, mid-periphery, and periphery of the cluster in area 18 as a function of age. Error bars represent S.E.M. 19 Bouton Density (10^6/mm^3) Overall Bouton Density in Area 19 16 12 core 8 mid-periphery 4 periphery 0 2 4 6 8 Age (weeks) 10 12 Figure 4b. Overall bouton density in the core, mid-periphery, and periphery of the cluster in area 19 as a function of age. Error bars represent S.E.M. Bouton Density (10^6/mm^3) Overall Bouton Density in Area 21 16 12 core 8 mid-periphery 4 periphery 0 2 4 6 8 Age (weeks) 10 12 Figure 4c. Overall bouton density in the core, mid-periphery, and periphery of the cluster in area 21 as a function of age. Error bars represent S.E.M. 20 Bouton Density (10^6/mm^3) Overall Bouton Density in Ssy 16 12 core 8 mid-periphery 4 periphery 0 2 4 6 8 Age (weeks) 10 12 Figure 4d. Overall bouton density in the core, mid-periphery, and periphery of the cluster in Ssy as a function of age. Error bars represent S.E.M. Developmental changes in interbouton intervals After observing developmental changes in bouton density, we decided to measure changes in interbouton intervals to find out if elimination of individual boutons along axons could account for the decline in bouton density with age. Figures 5a and 5b show the median interbouton intervals ± S.E.M. in the core and periphery of the clusters in areas 18, 19, 21, and Ssy in 4, 5, 6, 8, and 10 week olds. In both the core and the periphery, from 4 to 6 weeks, all of the interbouton intervals range from 0.785 to 1.14 µm. However, from 8 to 10 weeks, the median interbouton intervals are considerably larger, ranging from 1.1 to 1.53 µm. The major increase in interbouton intervals occurs from 6 to 8 weeks, with little change thereafter. We performed a Kruskal-Wallis test and found significant differences in interbouton intervals as a function of age in all areas, at the P=0.01 level. 21 Interbouton Intervals in the Core Interbouton Intervals (µm) 2 1.5 18 1 19 21 0.5 Ssy 0 2 4 6 8 Age (weeks) 10 12 Figure 5a. Median interbouton intervals in the core of the clusters in areas 18, 19, 21, and Ssy as a function of age. Error bars represent S.E.M. Interbouton Intervals in the Periphery Interbouton Intervals (µm) 2 1.5 18 1 19 21 0.5 Ssy 0 2 4 6 8 Age (weeks) 10 12 Figure 5b. Median interbouton intervals in the periphery of the clusters in areas 18, 19, 21, and Ssy as a function of age. Error bars represent S.E.M. Developmental changes in axon density 22 While the elimination of individual boutons could explain the developmental decline in bouton density, it is also possible that the elimination of axonal branches plays a role. Thus, we calculated axon densities at different ages to reveal the contribution that this pruning has on bouton density changes. Figures 6a through 6c display the axon density in the core, midperiphery, and periphery of each cluster in areas 18, 19, 21, and Ssy at 4, 5, 6, 8 and 10 weeks of age. The axon density in the core and mid-periphery are very similar, but that of the periphery is lower, particularly at the earlier ages. There is a lot of variability in the changes that occur with age among areas. For example, in the core of the cluster, the axon density of area 18 increases from 4 to 5 weeks, decreases from 5 to 8, but then increases slightly from 8 to 10. In contrast, the core density in area 19 decreases steadily from 4 to 8 weeks, but then increase from 8 to 10. The density in the core of both 21 and Ssy increases from 4 to 5 weeks, but shows a steady decline afterwards. Even within an area, there is variability in the trends of axon density with age in different parts of the cluster. As mentioned earlier, the density in the core of the Ssy cluster increases from 4 to 5 weeks and decreases thereafter. However, in the mid-periphery, the axon density in Ssy shows a steady decline from 4 to 8 weeks, but an increase from 8 to 10. In the periphery, however, Ssy displays very little change in axon density from 4 to 8 weeks, but a large decline from 8 to 10. Indeed, using a permutation test, we found that the only areas showing a significant change with age in all regions of the cluster were areas 21 and Ssy. It is difficult to make generalizations about the trends occurring in axon density with age, due to the high variability among and within areas. 23 Axon Density in the Core Axon Density (µm x10^6/mm^3) 14 12 10 8 18 6 19 21 4 Ssy 2 0 2 4 6 8 Age (weeks) 10 12 Figure 6a. Average axon density in the core of the clusters in areas 18, 19, 21, and Ssy as a function of age. Error bars represent S.E.M. Axon Density (µm x10^6/mm^3) Axon Density in the Mid-Periphery 14 12 10 8 18 6 19 4 21 Ssy 2 0 2 4 6 8 Age (weeks) 10 12 Figure 6b. Average axon density in the mid-periphery of the clusters in areas 18, 19, 21, and Ssy as a function of age. Error bars represent S.E.M. 24 Axon Density in the Periphery Axon Density(µm x10^6/mm^3) 14 12 10 8 18 19 6 21 4 Ssy 2 0 2 4 6 8 10 12 Age (weeks) Figure 6c. Average axon density in the periphery of the clusters in areas 18, 19, 21, and Ssy as a function of age. Error bars represent S.E.M. Developmental changes in neuronal density To correlate changes in bouton density with age to changes in neuronal density, we obtained this latter measurement in layers 2 to 6 in areas 17, 18, 19, 21, and Ssy at 4 and 6 weeks of age, as well as in adults. Figures 7a to 7e show the average neuronal density ± S.E.M. at 4 weeks, 6 weeks, and adulthood in each layer. Neuronal density decreased between the ages of 4 and 6 weeks in all layers in areas 17 and 18, though layer 5 of area 18 shows a more modest decline. From 6 weeks to adulthood, however, neuronal density did not change much in the majority of layers in these two areas. In Ssy, only layers 2 and 3 show a big decline in neuronal density in 6 week olds compared to 4 week olds. However, the density in all the layers in Ssy decreased by a much larger magnitude from 4 weeks of age to adulthood. Much of this decrease is accounted for by a decline in density between 6 weeks of age and adulthood. Surprisingly, neuronal density in all the layers of areas 19 and 21 remained essentially the same throughout 25 development. Despite the different magnitudes of changes in neuronal density occurring in each area and layer, a permutation test revealed that the only significant changes with age at the P<0.05 level occurred in layers 3-6 of area 17, layer 4 of area 18, and all layers in Ssy with the exception of layer 3. Neuroal Density (neurons x10^4/ mm^3) Neuronal Density in Layer 2 35 30 25 17 20 18 15 19 10 21 5 Ssy 0 4 weeks 6 weeks Age adults Figure 7a. Average neuronal density in layer 2 of areas 17, 18, 19, 21 and Ssy among the different age groups. Error bars represent S.E.M. 26 Neuroal Density (neurons x10^4/mm^3) Neuronal Density in Layer 3 35 30 25 17 20 18 15 19 10 21 5 Ssy 0 4 weeks 6 weeks Age adults Figure 7b. Average neuronal density in layer 3 of areas 17, 18, 19, 21 and Ssy among the different age groups. Error bars represent S.E.M. Neuroal Density (neurons x10^4/mm^3) Neuronal Density in Layer 4 35 30 25 17 20 18 15 19 10 21 5 Ssy 0 4 weeks 6 weeks Age adults Figure 7c. Average neuronal density in layer 4 of areas 17, 18, 19, 21 and Ssy among the different age groups. Error bars represent S.E.M. 27 Neuroal Density (neurons x10^4/mm^3) Neuronal Density in Layer 5 30 25 20 17 18 15 19 10 21 5 Ssy 0 4 weeks 6 weeks Age adults Figure 7d. Average neuronal density in layer 5 of areas 17, 18, 19, 21 and Ssy among the different age groups. Error bars represent S.E.M. Neuroal Density (neurons x10^4/mm^3) Neuronal Density in Layer 6 35 30 25 17 20 18 15 19 10 21 5 Ssy 0 4 weeks 6 weeks Age adults Figure 7e. Average neuronal density in layer 6 of areas 17, 18, 19, 21 and Ssy among the different age groups. Error bars represent S.E.M. 28 Figure 8a shows the average neuronal density ± S.E.M. in all the layers in area17 at different ages. Neuronal density decreased from 4 to 6 weeks in all layers, but did not change much from 6 weeks to adulthood. In the 4 week olds, layer 2 had the highest neuronal density (31.6 x 105 neurons/mm3) followed closely by layer 6 (30.4 x 104 neurons/mm3). At 21.6 x 104 neurons/mm3, the density of layer 5 was much lower than that of all the other layers at 4 weeks. At 6 weeks, there were no large differences between the densities of layers 2, 3, 4, or 5. However, layer 6 had a larger density than all the other layers. In adults, as in the 4 week olds, layer 2 contained the highest neuronal density (21.6 x 104 neurons/mm3), and layer 6 contained the second highest density (18.6 x 104 neurons/mm3). Neuronal Density (neurons x10^4/mm^3) Neuronal Density in Area 17 35 30 25 20 4 weeks 15 6 weeks 10 adult 5 0 2 3 4 Cortical Layer 5 6 Figure 8a. Average neuronal density in layers 2 to 6 in area 17 in the different age groups. Error bars represent S.E.M. Figure 8b shows the developmental changes in neuronal density in the 5 layers in area 18. Similar to the changes in area 17, a decrease in neuronal density from 4 to 6 week olds was observed in the layers of area 18. At 26.9 x 104 neurons/mm3, layer 2 contained the highest 29 density in the 4 week olds. From 6 weeks to adulthood, there was actually a large increase in neuronal density in layers 2 and 3. In adults, as in the 4 week olds, layer 2 contained the highest neuronal density (19.4 x 104 neurons/mm3). Layer 5 was the least dense layer at all ages: 12.5, 9.1, and 9.7 x 104 neurons/mm3 in the 4 week olds, 6 week olds, and adults, respectively. Neuronal Density (neurons x10^4/mm^3) Neuronal Density in Area 18 35 30 25 20 4 weeks 15 6 weeks 10 adults 5 0 2 3 4 Cortical Layer 5 6 Figure 8b. Average neuronal density in layers 2 to 6 in area 18 in the different age groups. Error bars represent the S.E.M. Figure 8c shows the average neuronal densities in layers 2 to 6 of area 19. All of the layers show very little change in density from 4 to 6 weeks, and from 6 weeks to adulthood. However, we do observe different densities among the layers within a given age. At all ages, layer 2 contains the highest density and layer 5 the lowest. At 12.5 and 15.6 x 104 neurons/mm3, layer 6 has the second highest density in the 4 and 6 week olds, respectively. However, in the 30 adults, layer 4 has the second highest density, though only slightly higher than that of layers 3 and 6. Neuronal Density (neurons x10^4/mm^3) Neuronal Density in Area 19 35 30 25 20 4 weeks 15 6 weeks 10 Adult 5 0 2 3 4 Cortical Layer 5 6 Figure 8c. Average neuronal density in layers 2 to 6 in area 19 in the different age groups. Error bars represent the S.E.M. Figure 8d shows that like area 19, area 21 does not show many changes in neuronal density with age. Layers 2 and 5 contain the highest and lowest densities at each age, respectively. Layers 3 and 4 contain very similar densities in all the age groups, ranging from 11.5 to 14.1 x 104 neurons/mm3. In layer 6, there is an increase in density from 4 to 6 weeks of age, from 10.3 to 16.0 x 104 neurons/mm3. However, in the adults, the density is between that of the 4 and 6 week olds, at 12.1 x 104 neurons/mm3. 31 Neuronal Density (neurons x10^4/mm^3) Neuronal Density in Area 21 35 30 25 20 4 weeks 15 6 weeks 10 Adult 5 0 2 3 4 Cortical Layer 5 6 Figure 8d. Average neuronal density in layers 2 to 6 in area 21 in the different age groups. Error bars represent the S.E.M. Figure 8e shows that in Suprasylvian, the major declines in density were observed when 4 week olds are compared to adults. This is in part due to the decline in neuronal density from 6 weeks of age to adulthood, during which layers 2, 4, and 5 show the biggest decline. At 4 weeks, the density in layer 2 was substantially higher than that in layers 3-6. In contrast, the density in layer 5 was the lowest at 15.1 x 104 neurons/mm3. The density in Layer 2 remained the highest in 6 week olds, while the density in layer 5 was only slightly lower than that of layers 2, 3, and 6 at this same age. In adults, layer 2 was still the highest, but the difference between this density and that of other layers was substantial only for layers 4 and 5. Layer 5, however, contained a much lower density than that of any other layer in adulthood. 32 Neuronal Density (neurons x10^4/mm^3) Neuronal Density in Ssy 35 30 25 20 4 weeks 15 6 weeks 10 adult 5 0 2 3 4 Cortical Layer 5 6 Figure 8e. Average neuronal density in layers 2 to 6 in Ssy in the different age groups. Error bars represent the S.E.M. Thalamo-cortical and cortico-cortical synapses We measured several characteristics of synapses in layer 4 of area 17, and we were able to use vglut2 to differentiate thalamo-cortical from cortico-cortical synapses. In vglut2+ profiles, we found that 13% of synapses were vglut2+. With the exception of one UC vglut2+ synapse, all of these thalamo-cortical synapses were asymmetric. Despite the different proportions and characterizations of thalamo-cortical versus cortico-cortical synapses, we did not find any differences in PSD across any of the comparisons we performed. Figure 9a shows that the PSD lengths of area 17 layer 4 synapses did not differ between vglut2+ and vglut2- profiles. Furthermore, there is no significant difference between the PSD lengths of asymmetric, symmetric, or UC synapses. We further compared the PSD lengths of all vglut2+ to that of all vglut2- synapses, as well as the PSD lengths of asymmetric vglut2+ to that of asymmetric vglut2- synapses, and as figure 9b shows, there were no differences. In all of our PSD length 33 comparisons we performed non-parametric Kruskal-Wallis tests (for 3-way comparisons) or Mann-Whitney tests (for pairwise comparisons) and found no significant differences. PSD lengths 0.4 Length (µm) 0.3 asym 0.2 sym UC 0.1 0 vglut2- vglut2+ Figure 9a. Average PSD lengths of asymmetric, symmetric, and uncharacterized synapses in layer 4 of area 17 in vglut2+ and vglut2- profiles. Error bars represent the S.E.M. asym=asymmetric, sym=symmetric, UC=uncharacterized. PSD lengths Length (µm) 0.40 0.30 0.20 vglut2+ vglut2- 0.10 0.00 all synapses asym synapses Figure 9b. Average PSD lengths of vglut2+ and vglut2- synapses in layer 4 of area 17. Error bars represent the S.E.M. 34 In addition to PSD lengths, we also calculated synaptic density in layer 4 of V1 to compare the strengths of synaptic input from different types of synapses. Figure 10 shows that the total synaptic density was 6.37 x 108 synapses per mm3. Of this density, approximately two thirds is due to asymmetric synapses. On the contrary, the density of UC synapses is less than one third of the total density, while the density of symmetric synapses is even smaller. Density (synapses x 10^8/mm^3 ) Synaptic Density in Layer 4 of V1 8 6 4 2 0 total asym UC sym Figure 10. Average synaptic density in layer 4 of V1. Error bars represent the S.E.M. Asym=asymmetric, sym=symmetric, and UC=uncharacterized. One last measure we calculated was the identity of the post-synaptic target in layer 4 of V1. We found that, on average, spines made up 69% of the targets, while dendrites made up 31%. Discussion Decline in feedforward projections One of the main findings of our study was the developmental decline of feedforward projections to extrastriate areas. We found that bouton density, a proxy for synapses, declines in the core, mid-periphery, and periphery of areas 18, 19, 21, and Ssy from 4 to 10 weeks of age. 35 As expected, when we calculated overall bouton density, we found lower densities than when we calculated peak bouton density. However, the developmental changes in bouton density, namely, a decline after 6 weeks of age, are similar in both types of measurements. Thus, our initial methodology for counting boutons does not appear to have masked any changes occurring with age. The observed declines in bouton density are in agreement with the phenomenon of exuberance, which posits that more axons and connections are produced than will persist in adulthood (Innocenti and Price, 2005). However, unlike the synaptic zinc decline that occurs right after eye opening (from 5 to 6 weeks), the major decline in bouton density in all areas occurs from 6 to 8 weeks of age. Nevertheless, just as Khalil and Levitt (2013) found that all areas reach their lowest intensity of synaptic zinc by the same age but decline by different amounts, we find that bouton density in all areas follows the same trends with age, implying a similar developmental timecourse for these different areas. At this moment, we cannot explain why is a decline in peak bouton density from 4 to 5 weeks of age, but an increase from 5 to 6. One reason that could explain these findings is the sample size, as we only had one 5 week old sample, while we had two samples at each of the other ages. In addition, the 5 week old case was one of the last cases from which we sampled. Thus, it is possible that our criteria for marking boutons changed. It is also probable that the particular 5 week old animal we used had a lower density than the average 5 week old and was not representative of the population. We calculated interbouton intervals to elucidate the mechanism of changes in bouton density, as an increase in interbouton intervals would imply that individual boutons are being eliminated. In addition, the interbouton interval data suggest a high degree of homogeneity in extrastriate areas and provide some insight on the arrangement of axons from the core to the periphery of each cluster. Though bouton density might vary initially in different areas, the 36 interbouton intervals in all areas are very similar from 4 to 6 weeks of age, during which they are approximately 1µm. After 6 weeks, the intervals increase by approximately 50%. Although the bouton density data are quite noisy, we do observe a decline by half in some areas from 4 to 10 weeks of age. However, since the interbouton intervals are not increasing by a factor of 2, we know that the increase in intervals only partially accounts for the decline in bouton density. Furthermore, we observed that bouton density is often, but not always, higher in the core than in the periphery, but the interbouton intervals are nearly identical. This finding suggests that there are more axons in the core than in the periphery, but the axons do not differ. In addition to the elimination of individual boutons, it is also possible that axonal branches are being pruned with age; to find out if this is the case, we measured axon densities. Unlike changes in bouton density and interbouton intervals, the trends we observe in axon density with age are much more variable between areas, and even between different parts of the cluster within a given area. All of the areas show a decline from 4 to 8 weeks in at least one part of the cluster. However, at 10 weeks, the axon density is higher than at 8 weeks in several areas in different parts of the cluster. This increased axon density at 10 weeks could be due to the case having a larger than average CTb injection core. This could have led to the labeling of more axons, and the resulting high density. A large variability we observe in axon density could also result from our using only one sample for this measure at each age. Furthermore, while we calculated bouton density and interbouton intervals by observing the brain section under 100x magnification, the axon density counts were done at 40x. As a result, the axon density counts are less accurate than the other two measures. Despite the noisiness of the data, we do see some modest declines in density with age, particularly in areas 21 and Ssy, which implies that axonal 37 branches are being eliminated. Thus, the decline in bouton density seems to be due to the elimination of individual boutons as well as pruning of axons. Developmental changes in neuronal density To further understand the anatomical changes occurring in the developing visual cortex, we decided to also measure neuronal density and observe if these changes mirror changes in bouton density. Indeed, we found that areas 17 and 18 show a decrease of about 50%, similar to the changes observed in bouton density. However, while the major decline in bouton density occurs after 6 weeks, neuronal density in areas 17 and 18 decreases between 4 and 6 weeks of age. Furthermore, while we find similar changes in bouton density in all areas, this is not the case when we consider neuronal density. Suprasylvian develops at a slower rate than areas 17 and 18, as the greatest decline in neuronal density in Ssy occurs between 6 weeks of age and adulthood. Interestingly, areas 19 and 21 do not show any major changes in density during development. We presume that most of the decline in neuronal density is likely due to the expansion of the brain with age, which might explain why areas 19 and 21 do not show a decline in neuronal density with age. Since these two areas are situated near the lateral sulcus, the brain might expand less in this area, resulting in fewer developmental changes in the density of these areas. Interestingly, if we focus on neuronal density in layer 4 of extrastriate areas, where we made the quantifications of feedforward bouton densities, we observe that the decline in neuronal density in this layer after 6 weeks does not fully account for the decline in bouton density. Thus, the decline in bouton density does not simply reflect the expansion of the brain that could be spacing the target neurons of feedforward projections farther apart. Instead, as stated earlier, changes in bouton density seem to reflect elimination of boutons and pruning of axons. 38 Thalamo-cortical versus cortico-cortical synapses While we can gain insight into the relative strengths of feedforward projections to different extrastriate cortex at different ages by calculating bouton densities, we need to study synapse at the EM level to understand the distribution and characteristics of different types of synapses. We began by investigating thalamo-cortical and cortico-cortical inputs in layer 4 of primary visual cortex in the adult. We found that only a small proportion of synapses in this layer, namely, 13%, comes from the LGN. This is consistent with what other researchers have found, as they identify that the majority of synapses in layer 4 of area 17 are cortico-cortical in nature (Ahmed et al., 1994). In addition, all vglut2+ synapses were characterized as asymmetric, except for one synapse which was en face, and thus, was listed as UC. Again, this is consistent with previous findings, as Nahmani and Erisir (2005) identified that vglut2 was localized in asymmetric, and thus, excitatory, synapses. However, while these researchers found that vglut2+ synapses have a larger PSD, we did not find any differences in PSD length when we compared vglut2+ to vglut2- synapses. Nor did we find any significant differences in the PSD length of asymmetric, symmetric, or UC synapses. A possible explanation for our lack of differences in PSD lengths could be the angle at which the synapses were cut. Since we are working with twodimensional images, we are not always measuring the largest extent of the PSD in the synapse, but rather, the extent that is visible in the particular image we obtain. Thus, measurements using three-dimensional reconstruction would be useful to validate or reject our findings. Despite our null findings of PSD length differences, we did find that each type of synapse makes differential contributions to the synaptic density in layer 4. In particular, the density of asymmetric synapses was substantially larger than that of symmetric or UC synapses. Furthermore, our data indicate 39 that synapses in layer 4 target spines more than dendrites, which is also commensurate with the results of Nahmani and Erisir (2005). Conclusions We are interested in understanding the development of the visual cortex due to the refinement it undergoes post-natally and its susceptibility to be influenced by visual experience. We have begun to elucidate some of the mechanisms that are at work in the developing visual cortex of the ferret. We found that the there is a decrease in the density of boutons that area 17 projects to areas 18, 19, 21, and Ssy between 4 and 10 weeks of age, and the changes are similar in all areas, suggesting a similar timecourse of refinement. This finding is the opposite of what we hypothesized, as we expected different areas to mature with different timecourses, likely elucidating a mechanism for the development of different visual functions. At these same ages, interbouton intervals increase with age, partly accounting for the decline in density, though elimination of axonal branches also seems to play a role. Our data suggest that the major changes in connectivity of feedforward projections occur after 6 weeks of age. In contrast, changes in neuronal density vary more by area, with areas 17 and 18 showing a decline before Ssy. On the other hand, areas 19 and 21 do not show any changes in neuronal density with age. The changes that we do observe in both bouton and neuronal density, however, occur during or shortly after eye opening in the ferret. This could indicate that visual experience plays a role in driving the refinement of these developmental changes. We are now beginning to quantify how characteristics and distributions of particular types of synapses in the visual cortex change with age. Our next step will be to use EM techniques and construct three-dimensional reconstructions 40 of synapses in different visual areas at different ages to have a much richer understanding of the refinement that the visual cortex undergoes. 41 References Ahmed, B., Anderson, J. C., Douglas, R. J., Martin, K. A. C., & Nelson, C. (1994). Polyneuronal innervation of spiny stellate neurons in cat visual cortex. Journal of Comparative Neurology, 341:39-49. Angelucci, A., Levitt, J. B., Walton, E. J. S., Hupé J., Bullier, J., & Lund, J. S. (2002). Circuits for local and global signal integration in primary visual cortex. Journal of Neuroscience, 22:8633-8646. Barnette, A. R., Neil, J. J., Kroenke, C. D., Griffith, J. L., Epstein, A. A., Bayly, P. V., Knutsen, A. K., and Inder, T. E. (2009). Characterization of brain development in the ferret via MRI. Pediatric Research, 66:80-84. Beul, S. F., & Hilgetag, C. C. (2015). Towards a “canonical” agranular cortical microcircuit. Frontiers in Neuroanatomy, 8:1-8. Binzegger, T., Douglas, R. J., & Martin, K. A. C. (2009). Topology and dynamics of the canonical circuit of cat V1. Neural Networks, 22:1071-1078. Binzegger, T., Douglas, R. J., & Martin, K. A. C. (2004). A quantitative map of the circuit of cat primary visual cortex. Journal of Neuroscience, 24:8441-8453. Bopp, R., Maçarico de Costa, N., Kampa, B. M., Martin, K. A. C., & Roth, M. M. (2014). Pyramidal cells make specific connections onto smooth (GABAergic) neurons in mouse visual cortex. PLoS Biology, 12:1-12. Braddick, O. J., Wattam-Bell, J., Atkinson, J. (1986). Orientation specific cortical responses develop in early infancy. Nature, 320:617–619. 42 Briones, T. L., Klintsova, A. Y., & Greenough, W. T. (2004). Stability of synaptic plasticity in the adult rat visual cortex induced by complex environment exposure. Brain Research, 1018:130-135. Budd, J. M. L. (1998). Extrastriate feedback to primary visual cortex in primates: a quantitative analysis of connectivity. The Royal Society, 265:1037-1044. Callaway E. M. (2004). Feedforward, feedback and inhibitory connections in primate visual cortex. Neural Networks, 17:625-632. Cantone, G., Xiao, J., McFarlane, N., & Levitt, J. B. (2005). Feedback connections to ferret striate cortex: Direct evidence for visuotopic convergence of feedback inputs. Journal of Comparative Neurology, 487:312-331. Costa, N. M., & Martin, K. A. C. (2010). Whose cortical column would that be? Frontiers in neuroanatomy, 4:1-10. DeFelipe, J., Marco, P., Busturia, I., & Merchán-Peréz, A. (1999). Estimation of the number of synapses in the cerebral cortex: Methodological considerations. Cerebral Cortex, 9:722732. Douglas, R. J., Martin, K. A. C., & Whitteridge, D. (1989). A canonical microcircuit for neocortex. Neural Computation, 1:480-488. Durack, J. C. & Katz, L. C. (1996). Development of horizontal projections in layer 2/3 of ferret visual cortex. Cerebral Cortex, 6:178-183. Garcia-Marin, V., Ahmed, T. H., Afzal, Y. C., & Hawken, M. J. (2012). Distribution of vesicular glutamate transporter 2 (VGlut2) in the primary visual cortex of the macaque and human. Journal of Comparative Neurology, 521:130-151. 43 Garcia-Marin, V., Blazquez-Llorca, L., Rodriguez, J. R., Boluda, S., Muntane, G., Ferrer, I., & DeFelipe, J. (2009). Diminished perisomatic GABAergic terminals on cortical neurons adjacent to amyloid plaques. Frontiers in Neuroanatomy, 3:1-18. Hupé, J. M., James, A. C., Payne, B. R., Lomber, S. G., Girard, P., & Bullier, J. (1998). Cortical feedback improves discrimination between figure and background by V1, V2, and V3 neurons. Nature, 394:784-787. Innocenti G. M. & Price D. J. (2005). Exuberance in the development of cortical networks. Nature Reviews Neuroscience, 6:955-965. Jackson, C. A., Peduzzi, J. D., & Hickey, T. L. (1989). Visual cortex development in the ferret. I. genesis and migration of visual cortical neurons. Journal for Neuroscience, 9:1242-1253. Khalil, R. & Levitt, J. B. (2013). Zinc histrochemistry reveals circuit refinement and distinguishes areas in the developing ferret cerebral cortex. Brain Structure & Function, 218:1293-1306. Ko, H., Mrsic-Flogel, T. D., & Hofer, S. B. (2014). Emergence of feature-specific connectivity in cortical microcircuits in the absence of visual experience. Journal of Neuroscience, 34:9812-9816. Krug, K., Akerman, C. J., & Thompson, I. D. (2001). Responses in neurons in neonatal cortex and thalamus to patterned visual stimulation through the naturally closed lids. Journal of Neurophysiology, 85:1436-1443. Lamme, V. A. F., Supèr, H., & Spekreijse, H. (1998). Feedforward, horizontal, and feedback processing in the visual cortex. Current Opinion in Neurobiology, 8:529-535. 44 Nahmani, M., & Erisir, A. (2005). VGlut2 immunohistochemistry identifies thalamocortical terminals in layer 4 of adult and developing visual cortex. Journal of Comparative Neurology, 484:458-473. Sloan, S. A., & Barres, B. A. (2014). Mechanisms of astrocyte development and their contributions to neurodevelopmental disorders. Current Opinion in Neurobiology, 27:7581. Smith, S. L. & Trachtenberg, J. T. (2007). Experience-dependent binocular competition in the visual cortex begins at eye opening. Nature, 10:370-375. Sur, M., & Leamy, C. A. (2001). Development and plasticity of cortical areas and networks. Nature Reviews Neuroscience, 2:251-262. Trachtenberg, J. T. & Stryker, M. P. (2001). Rapid anatomical plasticity of horizontal connections in the developing visual cortex. Journal of Neuroscience, 21:3476-3482. Wattam-Bell J. 1991. The development of motion-specific cortical responses in infants. Vision Res 31:287–297. 45