## Abstract

The effects of normal aging on morphologic and electrophysiologic properties of layer 3 pyramidal neurons in rhesus monkey primary visual cortex (V1) were assessed with whole-cell, patch-clamp recordings in in vitro slices. In another cohort of monkeys, the ultrastructure of synapses in the layers 2–3 neuropil of V1 was assessed using electron microscopy. Distal apical dendritic branching complexity was reduced in aged neurons, as was the total spine density, due to specific loss of mushroom spines from the apical tree and of thin spines from the basal tree. There was also an age-related decrease in the numerical density of symmetric and asymmetric synapses. In contrast to these structural changes, intrinsic membrane, action potential (AP), and excitatory and inhibitory synaptic current properties were the same in aged and young neurons. Computational modeling using morphologic reconstructions predicts that reduced dendritic complexity leads to lower attenuation of voltage outward from the soma (e.g., backpropagating APs) in aged neurons. Importantly, none of the variables that changed with age differed in neurons from cognitively impaired versus unimpaired aged monkeys. In summary, there are age-related alterations to the structural properties of V1 neurons, but these are not associated with significant electrophysiologic changes or with cognitive decline.

## Introduction

Multiple domains of visual function are impaired in a significant proportion of aged humans (for review, see Owsley 2011) and nonhuman primates (Schmolesky et al. 2000; Leventhal et al. 2003; Zhang et al. 2008; Fu et al. 2013). For example, age-related impairments in visual processing and acuity such as reduced contrast sensitivity (Elliott et al. 1990, 2009; Delahunt et al. 2008), shape discrimination (Owsley et al. 1981; Betts et al. 2007; Delahunt et al. 2008; Barnes et al. 2011), and motion perception (Trick and Silverman 1991; Norman et al. 2003, 2010) have been described. Previous studies indicate that retinal and thalamic visual processing are relatively spared with aging in awake behaving monkeys (Ahmad and Spear 1993; Spear 1993; Spear et al. 1994; Kim et al. 1996; Schmolesky et al. 2000), despite significant loss of optic nerve fibers (Sandell and Peters 2001). In contrast, signal processing in the primary visual cortex (V1) is altered in many ways. Single-unit recordings from V1 pyramidal cells during execution of visual tasks in aged rhesus monkeys have provided evidence that electrophysiologic changes in these neurons may underlie deficits in visual function. In aged monkeys, in which visual selectivity for orientation, motion and direction, and surround suppression are reduced, single-unit recordings reveal that these sensory deficits are accompanied by increased levels of spontaneous activity, decreased dynamic range, and decreased signal-to-noise ratios in V1 pyramidal cells (Schmolesky et al. 2000; Leventhal et al. 2003; Zhang et al. 2008; Fu et al. 2013). Both prolonged response latency (Wang et al. 2005; Yu et al. 2005) and increased trial-by-trial variability (Yu et al. 2005; Yang et al. 2009) during visual tasks have also been observed in aged monkeys. Most of these studies also reported increased excitability and reduced response selectivity of V1 neurons recorded in vivo. There is also evidence that, with age, the excitatory/inhibitory synaptic balance may be altered in V1, as application of γ-aminobutyric acid (GABA) and its agonists leads to a recovery of function on visual tasks (Leventhal et al. 2003). Taken together, these studies suggest that changes to the fundamental electrophysiologic properties of individual V1 pyramidal neurons may underlie global changes in visual function. Yet to date, structural and electrophysiologic properties, which may underlie age-related changes observed in vivo, have not been assessed in individual neurons.

It is well established that neither neurons nor glial cells are lost from the aging rhesus monkey area V1 (Vincent et al. 1989; Peters et al. 1997; Hof et al. 2000; Giannaris and Rosene 2012). Peters and coworkers have performed an extensive series of structural studies to assess sublethal changes to V1 neurons in young, middle-aged, and aged rhesus monkeys (for review, see Peters and Kemper 2012). During normal aging, increased myelin sheath thickness (Peters et al. 2001), increased paranode frequency (Peters and Sethares 2003), decreased layer 1 thickness and synapse density in V1 (Peters and Sethares 2002), and loss of fibers and increased dystrophic myelin in the splenium of the corpus callosum (unpublished data) have all been observed, but none of these changes correlate with cognitive decline. On the other hand, increased frequency of altered myelin sheaths and nerve fiber loss (Nielsen and Peters 2000; Peters et al. 2000) and also increased frequency of oligodendrocytes (Peters, Verderosa, et al. 2008) have been found to correlate with degree of cognitive decline in aged monkeys.

In the present study, we set out to investigate age-related changes to the structure of neurons and synapses in layer 3 of V1 and their potential functional consequences for the electrophysiologic response properties of individual pyramidal cells in the cortex. Previous work has shown marked changes with age in the morphologic and electrophysiologic properties of layer 3 (but not layer 5) pyramidal neurons in the dorsolateral prefrontal cortex (PFC) of the rhesus monkey, a higher-order association cortex (for review, see Luebke, Barbas, et al. 2010). Here, we assess whether similar changes occur in a primary sensory cortex, namely V1. The most marked structural change during normal aging in monkey PFC neurons is a reduction in dendritic spine density (Duan et al. 2003; Kabaso et al. 2009), likely due to a specific loss of thin spines (Dumitriu et al. 2010). Electron microscopic analyses of the numerical density of synapses in layers 2–3 of the rhesus monkey PFC reveal a loss of approximately 30% of synapses with age, with asymmetric (excitatory) and symmetric (inhibitory) synapses being lost at the same rate (Peters, Sethares, et al. 2008). Moreover, there is a strong correlation between cognitive impairment and the numerical density of asymmetric synapses in layers 2–3 of the PFC, and a weaker correlation between cognitive impairment and symmetric synapse loss (Peters, Sethares, et al. 2008). There are also functional changes with aging in layer 3 pyramidal neurons in the rhesus PFC, including markedly increased evoked action potential (AP) firing rates (Chang et al. 2005) and significantly reduced excitatory (EPSC) and increased inhibitory postsynaptic current (IPSC) frequency (Luebke et al. 2004). In the present study, both morphologic and electrophysiologic properties of individual layer 3 pyramidal cells of rhesus monkey V1 were examined to ascertain if these properties change with age and/or are associated with cognitive decline.

## Materials and Methods

### Experimental Subjects

All rhesus monkeys (Macaca mulatta) were part of a larger program of studies examining the impact of normal aging on the brain. A total of 9 young (8.5 ± 0.93 years old; range = 5–13 years old; 7 males, 2 females) and 8 aged (23.2 ± 0.67 years old; range = 21–26 years old; 5 females, 3 males) monkeys were used for the electrophysiologic and morphologic studies (Cohort 1). Of the 8 aged monkeys, 4 were classified as aged impaired (AI), because they had cognitive impairment index (CII; Herndon et al. 1997) scores >2.0, and 4 were classified as aged unimpaired (AU), because they had CII scores <2.0. Six young (6.8 ± 0.66 years old; range = 5–9 years old; 2 males, 4 females) and 9 aged (n = 27.9 ± 0.95 years old; range = 25–33 years old; all females) monkeys were used for the ultrastructural studies of synapses (Cohort 2). Of these 9 aged monkeys, 4 were classified as AI, because they had scores >2.0, and 4 were classified as AU, because they had CII scores <2.0. One aged monkey did not complete the full battery of behavioral tests, and therefore, cognitive data from this subject were excluded from the analyses.

Animals were housed individually in the Laboratory Animal Science Center (LASC) at the Boston University School of Medicine (BUSM) and kept under a 12-h light/dark cycle. Monkeys were initially obtained from the Yerkes National Primate Research Center at Emory University (Atlanta, GA, USA). Both the Yerkes National Primate Research Center and the BUSM LASC are fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, with animal research and maintenance conducted in strict accordance with the guidelines established by the NIH Guide for the Care and Use of Laboratory Animals and the US Public Health Service Policy on Humane Care and Use of Laboratory Animals.

### Cognitive Testing

Monkeys were tested on a battery of cognitive tasks similar to those used for testing human cognition. These include 3 visual recognition tests: the delayed nonmatch to sample (DNMS), the DNMS 2-min delay, and the delayed recognition span task, spatial and object. Each of these tests is designed to determine whether specific modalities of memory are affected by aging, and the tests were carried out as described by Moss and colleagues (Moss et al. 1988, 1997; Herndon et al. 1997; Moore et al. 2003, 2005, 2006). Overall cognitive status was quantified as the CII. The CII was derived based on the principal components analysis, which indicated that overall impairments were predicted by a weighted average of each subject's scores on the individual tasks (Herndon et al. 1997). To simplify this, the scores on each of the individual tasks for each subject were converted to a z-score relative to the mean performance of young adults, or the baseline reference group. The CII was then computed as a simple average of the 3 standardized scores, with positive numbers indicating increasing impairments in z-units from the mean of the baseline reference group of young adult monkeys, with a CII >2 standard deviations above the mean for young adult monkeys being defined as impaired. Visual function was not directly assessed in the monkeys used for these studies; however, structural magnetic resonance imaging scans of the entire brain including V1 and ophthalmologic examinations indicated preservation of V1 at a global level and of the peripheral visual apparatus in these subjects.

### Electrophysiologic and Morphologic Methods

#### Preparation of Slices and Whole-Cell Patch-Clamp Recordings

Monkeys were sacrificed as described previously (Chang et al. 2005; Luebke and Amatrudo 2010; Amatrudo et al. 2012). A 10-mm3 block of V1 was removed following perfusion of the brain with ice-cold Krebs-Henseleit buffer (concentrations, in mM: 6.4 Na2HPO4, 1.4 Na2PO4, 137 NaCl, 2.7 KCl, 5 glucose, 0.3 CaCl2, 1 MgCl2; pH 7.4, Sigma-Aldrich, St. Louis, MO, USA). The block was then cut into 300-µm thick coronal slices in ice-cold Ringer's solution (concentrations, in mM: 26 NaHCO3, 124 NaCl, 2 KCl, 3 KH2PO4, 2.5 CaCl2, 10 glucose, 1.3 MgCl2; pH 7.4; Sigma-Aldrich) on a vibrating microtome. After a 1-h equilibration period, individual slices were put into submersion-type recording chambers (Harvard Apparatus, Holliston, MA, USA) mounted on the stages of Nikon E600 infrared-differential interference contrast (IR-DIC) microscopes (Micro Video Instruments, Avon, MA, USA). Slices were continuously superfused with room temperature, oxygenated Ringer's solution during the entirety of the recording session (∼15 min per cell, 1–3 cells per slice) at a rate of 2–2.5 mL/min. Standard tight-seal, whole-cell patch-clamp recordings, and cell fillings were obtained from layer 3 pyramidal cells as described previously (Chang et al. 2005; Luebke and Amatrudo 2010; Amatrudo et al. 2012). Cells were visualized under infrared-differential interference optics, and electrodes were fabricated on a horizontal Flaming and Brown micropipette puller (Model P-87, Sutter Instruments, Novato, CA, USA). Electrodes contained a potassium methanesulfonate-based internal solution (concentrations, in mM: 122 KCH3SO3, 2 MgCl2, 5 EGTA, 10 NaHEPES, 1% biocytin; pH 7.4; Sigma-Aldrich), and had resistances of 3–6 MΩ in the external Ringer's solution. Data were acquired using EPC-9 or EPC-10 patch-clamp amplifiers (HEKA Elektronik, Lambrecht, Germany), using the PatchMaster acquisition software (HEKA Elektronik). Access resistance was monitored throughout the duration of each experiment, and signals were low-pass filtered at 10 kHz.

Passive membrane properties (resting membrane potential, input resistance, and membrane time constant), single AP properties (threshold, amplitude, duration at half maximal amplitude, rise and fall times), and repetitive AP firing properties were assessed as described previously (Chang et al. 2005; Luebke and Amatrudo 2010; Amatrudo et al. 2012). For assessment of spontaneous EPSC and IPSC, cells were recorded for 2 min at either −80 mV (for detection of spontaneous excitatory postsynaptic currents) or −40 mV (for detection of spontaneous IPSCs). Analysis was performed using the MiniAnalysis software (Synaptosoft, Decatur, HA, USA), with event detection threshold set at a maximum root mean square noise level (5 pA). Synaptic currents were assessed for their frequency, amplitude, area, and kinetics (rise and decay time constants, and event half-width).

#### Slice Processing, Confocal Imaging, and Preprocessing of Image Stacks

Following recording with simultaneous cell filling with biocytin, slices were fixed for 2 days at 4°C with 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS; pH 7.4). Following a PBS rinse, cells were placed in 0.1% Triton X-100 in PBS for 2 h at room temperature, and then for 2 days at 4°C in streptavidin–Alexa 488 (1 : 500 in PBS; InVitrogen, Carlsbad, CA, USA). Slices were mounted on slides and cover-slipped with Prolong Gold mounting medium (InVitrogen). Image stacks of fluorescently labeled cells were acquired using a Zeiss 510 confocal laser-scanning microscope. Fluorescence was emitted by Alexa-488 (Argon laser excitation) and collected using a 505-nm long-pass filter. Stacks were acquired at ×40 (1.5 digital zoom) using a Plan-Apochromat ×40/1.3 NA oil-immersion objective (210 µm working distance) at a resolution of 0.1 × 0.1 × 0.2 µm per voxel (153 µm2 field of view) for assessment of both dendritic and spine morphology. To determine dendritic diameter and further confirm spine subtype distributions, a second series of image stacks were acquired at ×100 (2.0 digital zoom) using a UPlan-FL ×100/1.3 NA oil-immersion objective (100 µm working distance) at a resolution of 0.022 × 0.022 × 0.1 µm per voxel (45 µm2 field of view). Images were acquired from one complete basilar dendritic branch (∼100 µm in length), the middle third of the main apical trunk (∼45 µm in length), and one distal apical dendritic branch (∼140 µm in length) of each neuron. To reduce signal blurring in the z-plane, each acquired stack of images was deconvolved using an Autodeblur software (Media Cybernetics, Bethesda, MD, USA). These deconvolved image stacks were then imported into the Volume Integration and Alignment System (VIAS) software (Rodriguez et al. 2003), aligned in 3-dimensions (3D) and integrated into a single volumetric dataset.

#### Morphometric Analyses of Somata, Axons, Dendrites, and Spines

The width and length (major diameter) of the somata of filled neurons were assessed using the RECONSTRUCT™ software (Fiala 2005). Volumetric datasets were imported into the AutoNeuron software (MBF Bioscience, Williston, VT, USA) and reconstructed in 3D. In most cells, at least 3000 µm of axons were filled and these were reconstructed in NeuronStudio, with boutons en passant and terminaux subsequently marked for determination of bouton density. For assessment of dendritic complexity (number of bifurcation nodes and intersections at a given distance from the soma) and dendritic length, these reconstructions were imported into the NeuroExplorer software (MBF Bioscience). These assessments were performed using a Sholl analysis, with concentric spheres placed at 20 µm increments and originating at the soma (Sholl 1953). A second Sholl analysis was performed, in which the apical and basal dendritic arbors were divided into proximal, middle, and distal thirds (Sholl 1953). Dendritic diameter was assessed from ×100 confocal images of middle and distal apical and of basilar branches using the Rayburst-based algorithm in the 64-bit version of NeuronStudio (Rodriguez et al. 2003, 2006; Wearne et al. 2005; available at http://www.mssm.edu/cnic), which automatically traced dendritic segments and generated a .swc file. This .swc file was imported into L-measure (Scorcioni et al. 2008; available at http://cng.gmu.edu:8080/Lm/), which extracted the mean dendritic diameter data. The reconstructions are available at NeuroMorpho.org.

For assessment of dendritic spine density (number of spines/µm) and distribution, integrated volumetric datasets were imported into NeuronStudio. Dendritic spines were automatically detected on previously generated dendritic reconstructions (Rodriguez et al. 2008). The accuracy of this automatic detection was manually checked by an operator. The maximal widths of spine heads (the widest part of the spine head perpendicular to the neck) and the lengths of spines (the maximum distance of the center of spine heads to the dendritic shaft) were measured automatically and cross-checked by an operator. Spines were classified as thin, mushroom, stubby, or filopodia based on spine head and neck morphometrics. Those lacking a neck were classified as stubby, while those with a maximal head width of >0.6 µm were classified as mushroom. Spines with a maximal head width of <0.6 µm were classified as either thin or filopodia, with thin spines having a neck length of ≤3 µm and filopodia having a neck length of >3 µm.

#### Cell Inclusion Criteria

For inclusion in electrophysiologic analyses, cells were required to have a resting membrane potential less than or equal to −55 mV, stable access resistance, an AP overshoot, and the ability to fire repetitive APs in response to prolonged depolarizing current steps. For inclusion in morphologic analyses, the criteria were: an intact soma, completely filled dendritic arbors and no cut dendrites in the proximal third of the apical dendritic arbor. The electrophysiologic analyses were performed on 2–6 pyramidal cells per monkey, while morphometric analyses were performed on 1–4 pyramidal cells per monkey.

### Electron Microscopy Methods

#### Tissue Samples and Processing

The morphology and distribution of synapses in layers 2–3 of V1 in all 15 of the young and aged monkeys in Cohort 2 were assessed using electron microscopy (EM). Animals used for ultrastructural studies were perfused with 1% paraformaldehyde and 1.25% glutaraldehyde in 0.1 M cacodylate or phosphate buffer (pH 7.4) as described previously (Peters et al. 2000, 2001; Peters, Sethares, et al. 2008; Peters, Verderosa, et al. 2008). Upon removal, one-half of the brain was then immersed in a stronger aldehyde solution, containing 2% paraformaldehyde and 2.5% glutaraldehyde in the same buffer used for the fixation for at least 1 week at 4°C. Small blocks of tissues were taken from V1, osmicated and embedded in araldite resin, as described previously (Peters et al. 2000, 2001; Peters, Sethares, et al. 2008; Peters, Verderosa, et al. 2008). Semithick (1 μm) sections were cut from the araldite-embedded blocks and oriented, so that the plane of section was parallel to the vertical lengths of the apical dendrites, and included the entire depth of V1 from the pia to the white matter. These semithick sections were mounted on glass slides and stained with toluidine blue for light microscopic examination. Layers 2–3 were demarcated using a light microscope equipped with a camera lucida. Thin sections (0.05 μm) were then taken from the same blocks used to produce thick sections and cut in the same vertical plan. These thin sections were mounted on copper mesh grids, stained with uranyl acetate and lead citrate, and then examined with a JEOL 100S transmission electron microscope. First, low magnification was used to draw the thin section to determine the location of layer 3. This was done by comparing the drawings of the sections from the light microscope and electron microscope. The sampling region in layers 2–3 was demarcated beginning approximately 100 μm deep to the pia and extending about 300 μm towards the white matter. The average block face area of V1 layers 2–3 sampled in each monkey was about 740 × 360 μm for young and 738 × 394 μm for aged animals. Within this area of neuropil in each monkey, 10–22 micrographs (area of each = 185 μm2) were systematically sampled and photographed at a magnification of ×6000. In taking the electron micrographs of layers 2–3, the cell bodies of neurons and neuroglial cells were avoided. The result was that the postsynaptic elements examined were either dendrites or dendritic spines, but not neuronal soma. EM negatives were then digitally scanned and analyzed using the RECONSTRUCT™ software for counting, tracing, and measuring objects from images of sections.

#### Synapse Counts and Measurement of Synaptic Profile Lengths

The methods and criteria employed in a previous study (Peters, Sethares, et al. 2008) were used here to count profiles of synaptic junctions and to measure the lengths of the postsynaptic densities. The profiles were identified as one of the following types of synapse: Asymmetric axospinous, asymmetric axodendritic, symmetric axospinous, symmetric axodendritic, or uncharacterized. A synapse was counted only if the presynaptic component contained at least 2 vesicles, both pre- and postsynaptic elements were evident and a postsynaptic density was apparent (Peters, Sethares, et al. 2008).

The majority of synaptic profiles can be easily identified as symmetric or asymmetric, based on at least 2 of the 3 classic criteria: Width of the synaptic cleft, shape of the synaptic vesicles, and thickness of the postsynaptic density (Peters et al. 1991). In rare cases (∼1% of the synapses encountered), the plane of section passed parallel to the synaptic junctions revealing en face synapses, and such profiles were not included in the synapse counts. When it was not possible to define the postsynaptic element of a given synapse, the synaptic profile was defined as uncharacterized. As an estimate of synapse size, the lengths of synaptic densities of the synaptic junctions were measured in RECONSTRUCT™, but only if there was a discernable cleft separating the pre- and postsynaptic elements. If the profile of a synaptic junction was curved or perforated, the length of the postsynaptic density was measured linearly between its 2 endpoints. A minimum of 100 asymmetric and 50 symmetric synaptic densities were measured per animal and the mean lengths determined. This method resulted in a total of 3357 synapses counted from a total volume of 860 μm3 from 6 young animals and in a total of 3299 synapses counted from a total volume of 1027 μm3 from 9 aged animals, with about 6.6% of the synapses uncharacterized with regard to their postsynaptic site. From the subset of characterized synapses, lengths of synaptic profiles were measured from a total 699 asymmetric and 302 symmetric synapses from 6 young animals and from a total of 923 asymmetric and 380 symmetric synapses from 9 aged animals. In each animal, synapses were counted from multiple sampling sites until the standard error of the mean (SEM) across sampling sites was <10% (average SEM across sampling sites was 4.4% for young animals, and 6.5% for aged animals).

#### Estimating Numerical Density of Synapses

To determine the numerical density of synapses, the empirical formula suggested by Colonnier and Beaulieu (1985), known as the size-frequency method, was used. The formula is NV = NA/d, where NV is the frequency of synapses per unit volume, NA is the number of synaptic profiles per unit area of electron micrographs, and d is the mean length of synaptic densities. The alternative method used to estimate synaptic density is the disector method, which involves preparing serial thin sections and examining identical fields in adjacent sections of a series of thin sections (e.g., Sterio 1984; Calverley et al. 1988). The Colonnier and Beaulieu size-frequency approach is inherently biased due to variability in size, shape, and orientation of synapses counted in single sections. Despite this, we chose to employ this method rather than the unbiased disector method for 4 reasons: (1) It enabled the comparison of findings in V1 to our previous work done in the PFC in the same animals (Peters, Sethares, et al. 2008); (2) it enabled us to extend and adequately distribute sampling sites across a large surface area of layers 2–3 of V1, which has heterogeneous functional subdomains (reviewed in Hirsch and Martinez 2006); (3) it enabled maximization of sampling across subjects, and (4) it is time efficient. Furthermore, the size-frequency and the disector approaches have been validated against each other and shown to yield comparable results (DeFelipe et al. 1999; Peters et al. 2001).

### Computational Modeling

The NEURON 7.1 simulation environment (Carnevale and Hines 2006) was used to model the attenuation of voltage outward from and inward towards the soma of young and aged V1 neurons (Brown et al. 1992; Carnevale et al. 1997). The mean outward and inward attenuations, $L¯out$ and $L¯in$, over all paths of the dendritic arbor were calculated as described previously (Kabaso et al. 2009; Amatrudo et al. 2012; Accession number 144553 on ModelDB, https://senselab.med.yal.edu/ModelDB) using sinusoidal inputs varying from 0 (DC) to 500 Hz and a time step of 0.05 ms. Digital reconstructions of 8 young and 7 aged neurons were imported into NEURON using the Import3D tool. All model neuron somata had the same dimensions (length 10.86 µm and diameter 12.96 µm), equal to the mean dimensions of all somata measured. In some simulations, dendritic spines were added uniformly to each neuron with density equal to the mean of the young and aged groups. As in Amatrudo et al. (2012), the surface area of a typical dendritic spine was estimated from a previous study (Hao et al. 2006), and the total dendritic surface area was added using the algorithm of Stratford et al. (1989). The size of each compartment was chosen to be less than 1/10th of the electrotonic length constant at 100 Hz. Passive membrane parameters of Cm = 1 µF/cm2, Ra = 150 Ω cm, and Rm = 20 833 Ω cm2 were taken from Amatrudo et al. (2012), where they were fit directly to electrophysiologic data from a young V1 neuron.

### Statistical Analysis

A 2-tailed Student's t-test was used for between-group comparisons. Relationships between variables were determined using a Pearson product–moment correlation. Significance was set at α = 0.05 for all statistical tests. All data are reported as means ± SEM.

## Results

### Aged Monkeys Are Segregated Into Impaired and Unimpaired Groups

A major goal of this study was to determine whether any age-related changes observed in V1 correlated with degree of cognitive impairment in aged subjects, as has been observed for other cortical areas such as the dorsolateral PFC (for review, see Luebke, Barbas, et al. 2010). For this purpose, all subjects were comprehensively assessed on a battery of cognitive tasks and assigned a CII score. There were no statistically significant differences in the data between male and female monkeys, so data from both sexes within each age cohort were pooled for statistical analyses of age effects. In both Cohort 1 (used for in vitro slice studies) and Cohort 2 (used for ultrastructural studies), 4 of 8 aged subjects had CII z-scores >2 and were thus classified as aged cognitively impaired (AI) and 4 had CII z-scores <2 and were thus classified as aged unimpaired (AU). This distribution enabled us to compare AI with AU subjects and young with aged, and also to perform linear correlations between variables that changed with age versus cognitive performance. Table 1 data indicate that no statistically significant difference in CII was seen between young and aged monkeys in either Cohort 1 (P = 0.30) or Cohort 2 (P = 0.09) nor between AI and AU subjects in Cohort 1 (P = 0.22) or Cohort 2 (P = 0.18). The simple explanation for this unexpected statistical finding is that, in both cohorts, there was one aged monkey in the impaired group that exhibited an unusually high degree of impairment (CII near 12), which caused the variability in the aged and aged impaired (AI) groups to be high. When these outliers were removed, P-values for aged versus young CII were <0.04 in both cohorts, and those for AI versus AU were <0.001 in both cohorts.

Table 1

Experimental subjects

Young Aged P-value AU AI P-value
Cohort 1
Age (years) 8.50 ± 0.93 23.21 ± 0.67 <0.0001 24.25 ± 0.99 22.18 ± 0.92 0.13
CII 1.00 ± 0.18 3.03 ± 1.29 0.30 1.37 ± 0.05 3.06 ± 2.40 0.22
DNMS—trials 457 ± 33.8 900 ± 395 0.46 404 ± 92 956 ± 714 0.24
DNMS—errors 98 ± 6.0 225 ± 107 0.43 89.3 ± 28 252 ± 190 0.23
DNMS 2 min 14.7 ± 3.92 10.9 ± 4.92 0.64 14.2 ± 8.90 9.84 ± 3.13 0.54
DRST spatial 2.57 ± 0.12 2.31 ± 0.12 0.24 2.22 ± 0.08 1.03 ± 0.62 0.53
DRST object 2.81 ± 0.09 2.80 ± 0.17 0.97 2.81 ± 0.18 1.25 ± 0.71 0.95
Cohort 2
Age (years) 6.83 ± 0.66 28.33 ± 1.01 <0.0001 27.0 ± 0.94 29.40 ± 1.68 0.24
CII 1.00 ± 0.53 2.17 ± 0.40 0.09 1.57 ± 0.23 2.65 ± 0.68 0.18
DNMS—trials 324 ± 75.1 545 ± 120 0.20 315 ± 81.4 729 ± 173 0.06
DNMS—errors 79.8 ± 18.4 157.3 ± 28.7 0.07 102.5 ± 28.4 201 ± 38.8 0.06
DNMS 2 min 21.0 ± 4.85 21.0 ± 2.43 1.00 20.3 ± 5.31 21.8 ± 2.50 0.78
DRST spatial 2.24 ± 0.05 2.05 ± 0.09 0.13 2.04 ± 0.11 2.06 ± 0.15 0.91
DRST object 3.54 ± 0.48 2.77 ± 0.14 0.29 nd 2.77 ± 0.20 nd
Young Aged P-value AU AI P-value
Cohort 1
Age (years) 8.50 ± 0.93 23.21 ± 0.67 <0.0001 24.25 ± 0.99 22.18 ± 0.92 0.13
CII 1.00 ± 0.18 3.03 ± 1.29 0.30 1.37 ± 0.05 3.06 ± 2.40 0.22
DNMS—trials 457 ± 33.8 900 ± 395 0.46 404 ± 92 956 ± 714 0.24
DNMS—errors 98 ± 6.0 225 ± 107 0.43 89.3 ± 28 252 ± 190 0.23
DNMS 2 min 14.7 ± 3.92 10.9 ± 4.92 0.64 14.2 ± 8.90 9.84 ± 3.13 0.54
DRST spatial 2.57 ± 0.12 2.31 ± 0.12 0.24 2.22 ± 0.08 1.03 ± 0.62 0.53
DRST object 2.81 ± 0.09 2.80 ± 0.17 0.97 2.81 ± 0.18 1.25 ± 0.71 0.95
Cohort 2
Age (years) 6.83 ± 0.66 28.33 ± 1.01 <0.0001 27.0 ± 0.94 29.40 ± 1.68 0.24
CII 1.00 ± 0.53 2.17 ± 0.40 0.09 1.57 ± 0.23 2.65 ± 0.68 0.18
DNMS—trials 324 ± 75.1 545 ± 120 0.20 315 ± 81.4 729 ± 173 0.06
DNMS—errors 79.8 ± 18.4 157.3 ± 28.7 0.07 102.5 ± 28.4 201 ± 38.8 0.06
DNMS 2 min 21.0 ± 4.85 21.0 ± 2.43 1.00 20.3 ± 5.31 21.8 ± 2.50 0.78
DRST spatial 2.24 ± 0.05 2.05 ± 0.09 0.13 2.04 ± 0.11 2.06 ± 0.15 0.91
DRST object 3.54 ± 0.48 2.77 ± 0.14 0.29 nd 2.77 ± 0.20 nd

P-values that are statistically significant (less than 0.05) are indicated in italics.

CII, cognitive impairment index; DNMS, delayed nonmatch to sample; DRST, delayed recognition span task; nd, not determined.

### Features of the Somata, Dendrites, and Axons of Aged Layer 3 Pyramidal Cells

Three-dimensional reconstructions of young and aged pyramidal neurons (Fig. 1A,B) were used for assessment of soma dimensions, dendritic diameter and topology, and axonal bouton density (Table 2). The mean length and width of the neuronal somata did not differ between pyramidal neurons from young and aged subjects (Table 2). The total arbor, total apical, and total basal dendritic lengths and the diameters of the mid-apical, distal apical, and basilar arbors were also the same in pyramidal neurons from the 2 age cohorts (Table 2). Finally, the density of total, en passant, and terminaux axonal boutons quantified on reconstructed axons were not significantly different in young versus aged neurons (Table 2).

Table 2

Morphologic properties

Young Aged P-value
Soma
Length (μm) 12.94 ± 0.38 12.97 ± 0.40 0.96
Width (μm) 11.03 ± 0.38 10.7 ± 0.31 0.49
Dendritic length
Total (μm) 3405 ± 333 2802 ± 203 0.13
Apical (μm) 1657 ± 180 1237 ± 112 0.052
Basal (μm) 1748 ± 169 1566 ± 119 0.37
Dendritic diameter
Distal apical (μm) 0.37 ± 0.06 0.19 ± 0.09 0.08
Mid-apical (μm) 1.03 ± 0.10 0.51 ± 0.24 0.052
Basal (μm) 0.33 ± 0.03 0.29 ± 0.09 0.61
Axon
Length reconstructed (μm) 3720 ± 737 3626 ± 548 0.91
Total bouton density (#/μm) 0.05 ± 0.01 0.06 ± 0.01 0.55
En passant density (#/μm) 0.028 ± 0.006 0.031 ± 0.004 0.65
Terminaux density (#/μm) 0.022 ± 0.006 0.026 ± 0.004 0.53
Young Aged P-value
Soma
Length (μm) 12.94 ± 0.38 12.97 ± 0.40 0.96
Width (μm) 11.03 ± 0.38 10.7 ± 0.31 0.49
Dendritic length
Total (μm) 3405 ± 333 2802 ± 203 0.13
Apical (μm) 1657 ± 180 1237 ± 112 0.052
Basal (μm) 1748 ± 169 1566 ± 119 0.37
Dendritic diameter
Distal apical (μm) 0.37 ± 0.06 0.19 ± 0.09 0.08
Mid-apical (μm) 1.03 ± 0.10 0.51 ± 0.24 0.052
Basal (μm) 0.33 ± 0.03 0.29 ± 0.09 0.61
Axon
Length reconstructed (μm) 3720 ± 737 3626 ± 548 0.91
Total bouton density (#/μm) 0.05 ± 0.01 0.06 ± 0.01 0.55
En passant density (#/μm) 0.028 ± 0.006 0.031 ± 0.004 0.65
Terminaux density (#/μm) 0.022 ± 0.006 0.026 ± 0.004 0.53
Figure 1.

Regression of the distal apical arbor in aged neurons. (A) NeuronStudio reconstructions of representative young and aged layer 3 pyramidal neurons. (B) Reconstructions of representative young and aged neurons used for dendritic analyses. (C) Left: Sholl analyses of the mean lengths, intersections, and nodes of the apical dendritic arbor; right: mean lengths, intersections, and nodes of the proximal, middle, and distal thirds of the apical dendritic arbor. *P < 0.05. Scale bars: A = 10 μm, B = 100 μm.

Figure 1.

Regression of the distal apical arbor in aged neurons. (A) NeuronStudio reconstructions of representative young and aged layer 3 pyramidal neurons. (B) Reconstructions of representative young and aged neurons used for dendritic analyses. (C) Left: Sholl analyses of the mean lengths, intersections, and nodes of the apical dendritic arbor; right: mean lengths, intersections, and nodes of the proximal, middle, and distal thirds of the apical dendritic arbor. *P < 0.05. Scale bars: A = 10 μm, B = 100 μm.

### Reduction in Complexity of the Distal Apical Arbor of Aged Layer 3 Pyramidal Cells

Dendritic branching topology was assessed with a Sholl analysis using concentric spheres separated by 20 µm. For each cell, the dendritic length between consecutive Sholl rings, the number of dendritic intersections at each ring, and the number of nodes between consecutive rings was determined. The lengths of dendritic segments between consecutive rings were significantly reduced in aged compared with young neurons at distances >200 µm from the soma (Fig. 1C, top left panel; P < 0.05), and aged neurons also exhibited fewer dendritic intersections at distances >180 µm from the soma (Fig. 1C, middle left panel; P < 0.05). The number of nodes (branch points) did not differ in the apical dendritic arbors of cells from the 2 age groups (Fig. 1C, bottom left panel). In contrast to the regression seen in the distal apical arbor of the aged neurons, there was no difference in the segment lengths, number of intersections, or number of nodes in the basal arbors of cells from the 2 age groups. To normalize for variability in the soma to pial surface distance of the neurons, apical and basal dendritic arbors were divided into proximal, middle, and distal thirds. The results from this normalization (Fig. 1C, right panels) mirror those seen with the Sholl analysis (Fig. 1C, left panels) in that reductions in dendritic lengths and intersections occurred in the distal portion of the apical tree (P < 0.05). When neurons from AI monkeys were compared with those from AU monkeys, no difference in any morphologic parameter was found.

### Dendritic Spine Density Is Lower, and the Proportions of Spine Subtypes Are Altered with Age

Dendritic spines were quantified along the entire extent of dendritic trees of young and aged V1 neurons (Figs 2 and 3). Mean spine density was reduced by approximately 24% across the total dendritic arbor of aged versus young neurons (Fig. 2B, top; P = 0.001), resulting from loss of spines along both apical and basal dendrites of aged neurons (Fig. 2B, bottom; apical, P = 0.023 and basal, P < 0.001). Interestingly when separated by the subtype, with aging, there were significantly fewer mushroom spines in the apical arbor (P = 0.024) and thin spines in the basal arbor (Fig. 2C; P = 0.002). There were no other major differences in density among other subtypes of either arbor between the 2 groups (Fig. 2C).

Figure 2.

Reduced density of dendritic spines in aged neurons. (A) ×100 confocal images of representative apical dendritic segments from young and aged neurons. Note the presence of thin, mushroom, stubby, and filopodia spine subtypes. (B) Bar graphs of mean total (top) and apical and basal (bottom) dendritic spine density in young versus aged neurons. (C) Bar graphs showing mean density of thin, mushroom, stubby, and filopodia spine subtypes in the apical (top) and basal (bottom) dendritic arbors. *P < 0.05; **P < 0.03. Scale bar in A = 2.5 µm.

Figure 2.

Reduced density of dendritic spines in aged neurons. (A) ×100 confocal images of representative apical dendritic segments from young and aged neurons. Note the presence of thin, mushroom, stubby, and filopodia spine subtypes. (B) Bar graphs of mean total (top) and apical and basal (bottom) dendritic spine density in young versus aged neurons. (C) Bar graphs showing mean density of thin, mushroom, stubby, and filopodia spine subtypes in the apical (top) and basal (bottom) dendritic arbors. *P < 0.05; **P < 0.03. Scale bar in A = 2.5 µm.

Figure 3.

Sholl analyses reveal uniform loss of spines across the apical and basilar dendritic arbors of aged neurons. (A) Representative young and aged layer 3 pyramidal neurons with color-coded spines superimposed on the dendritic reconstructions (blue = mushroom; yellow = thin; red = stubby; and gray = filopodia). (B) Sholl plots of the distribution of spine subtypes across the apical (4 left graphs) and basal (4 right graphs) arbors. Scale bar in A = 50 µm.

Figure 3.

Sholl analyses reveal uniform loss of spines across the apical and basilar dendritic arbors of aged neurons. (A) Representative young and aged layer 3 pyramidal neurons with color-coded spines superimposed on the dendritic reconstructions (blue = mushroom; yellow = thin; red = stubby; and gray = filopodia). (B) Sholl plots of the distribution of spine subtypes across the apical (4 left graphs) and basal (4 right graphs) arbors. Scale bar in A = 50 µm.

Figure 3A demonstrates qualitatively the distribution patterns of each spine subtype (thin, mushroom, stubby, and filopodia) across dendritic trees of a representative young and aged V1 neuron. To quantitatively assess whether proximal-to-distal distribution patterns of spine subtypes were altered with age, Sholl analyses were performed in which apical and basal arbors were separately subdivided using concentric spheres originating from the soma. Aged neurons had significantly fewer mushroom spines than young neurons within the majority of apical divisions of <200 µm from the soma (Fig. 3B; 40 µm, P = 0.049; 80 µm, P = 0.002; 100 µm, P = 0.031; 120 µm, P = 0.022; 140 µm, P = 0.002; 180 µm, P = 0.045). Thus, with age, mushroom spines were lost largely within the proximal two-thirds of the apical arbor (Fig. 3B). Similarly, spine loss in the basal arbor was restricted to the proximal two-thirds, where aged V1 neurons possessed fewer thin spines than young V1 neurons (40 µm from the soma, P = 0.006; 60 µm, P = 0.006). Finally, the mean maximal width and length of spines, assessed using the Rayburst algorithm of NeuronStudio, did not differ between young and aged neurons for any spine subtype or for all spines as a group (P > 0.06; young neurons: spine head width = 0.43 ± 0.02, spine length = 1.1 ± 0.05 µm; aged: spine head width = 0.46 ± 0.01 µm, spine length = 1.2 ± 0.08 µm).

### Synapse Densities Are Reduced While Synaptic Profile Lengths Are Largely Unchanged with Aging in V1

To complement the confocal imaging analyses of dendrites and spines described above, the distribution and morphology of synapses in the neuropil of layers 2–3 of V1 were examined in a second cohort of behaviorally characterized young and aged monkeys using EM (Fig. 4). The relative proportion of asymmetric and symmetric synapses was similar in young and aged monkeys (P = 0.06), with the majority of synapses being asymmetric (Fig. 5A). Among asymmetric synapses, the majority (young 74 ± 1%; aged 71 ± 2%) were axospinous, and the remainder were axodendritic (Figs 4A,B and 5B). These asymmetric synapse proportions were similar in both young and aged monkeys (P = 0.39). Among symmetric synapses, the axospinous and axodendritic synapses were similar in number (Figs 4A,C,D and 5B). In young monkeys, 44 ± 3% of symmetric synapses were axospinous and 56 ± 3% were axodendritic, a distribution pattern which is reversed in aged monkeys (Fig. 5B; 54 ± 3% axospinous, 46 ± 3% axodendritic; P = 0.052).

Figure 4.

Electron micrographs of the neuropil in layers 2–3 of V1 showing examples of synapses. (A) The field contains an axon terminal (A) forming an asymmetric synapse (black arrow) with a spine (sp) and containing round vesicles. In comparison, 3 nearby axon terminals (S1, S2, and S3) contain flat vesicles and form symmetric synapses (open arrows), with 2 terminals (S1 and S2) each forming a synapse on one dendrite (d1 and d2), and the other terminal (S3) forming 2 synapses on 2 dendrites (d3). (B) The field shows examples of asymmetric synapses (black arrows) formed by axon terminals containing round vesicles: 3 axon terminals (A1) each form an asymmetric synapse with one dendrite (d1). One axon terminal (A2) forms an asymmetric synapse with a spine (sp2). (C) The field shows an example of an axon terminal (S) forming a symmetric synapse (black arrow) on a spine (sp). (D) The field shows a spine (sp) forming 2 synapses (arrows) with 2 axon terminals, one synapse is asymmetric (A, black arrow) and the other symmetric (S, open arrow).

Figure 4.

Electron micrographs of the neuropil in layers 2–3 of V1 showing examples of synapses. (A) The field contains an axon terminal (A) forming an asymmetric synapse (black arrow) with a spine (sp) and containing round vesicles. In comparison, 3 nearby axon terminals (S1, S2, and S3) contain flat vesicles and form symmetric synapses (open arrows), with 2 terminals (S1 and S2) each forming a synapse on one dendrite (d1 and d2), and the other terminal (S3) forming 2 synapses on 2 dendrites (d3). (B) The field shows examples of asymmetric synapses (black arrows) formed by axon terminals containing round vesicles: 3 axon terminals (A1) each form an asymmetric synapse with one dendrite (d1). One axon terminal (A2) forms an asymmetric synapse with a spine (sp2). (C) The field shows an example of an axon terminal (S) forming a symmetric synapse (black arrow) on a spine (sp). (D) The field shows a spine (sp) forming 2 synapses (arrows) with 2 axon terminals, one synapse is asymmetric (A, black arrow) and the other symmetric (S, open arrow).

Figure 5.

Percentage distribution and numerical density of synapses in the neuropil in layers 2–3 of V1 of young and aged monkeys. (A) Relative proportion of asymmetric and symmetric synapses in young and aged monkeys. (B) Proportion of asymmetric (left panel) and symmetric (right panel) synapses, which are formed with spines (axospinous) and dendrites (axodendritic). (C) Mean number of asymmetric and symmetric synapses per mm3 in young and aged monkeys (**P < 0.001). (D) A plot of the number of asymmetric and symmetric synapses per mm3 in the neuropil of each animal against age.

Figure 5.

Percentage distribution and numerical density of synapses in the neuropil in layers 2–3 of V1 of young and aged monkeys. (A) Relative proportion of asymmetric and symmetric synapses in young and aged monkeys. (B) Proportion of asymmetric (left panel) and symmetric (right panel) synapses, which are formed with spines (axospinous) and dendrites (axodendritic). (C) Mean number of asymmetric and symmetric synapses per mm3 in young and aged monkeys (**P < 0.001). (D) A plot of the number of asymmetric and symmetric synapses per mm3 in the neuropil of each animal against age.

The mean density of all synapses in the neuropil of layers 2–3 of V1 per unit volume (NV) was lower in aged versus young monkeys (P = 0.002; Table 3). The reduction in overall synapse density is principally due to an approximately 43% loss of symmetric synapses with age (P = 0.0003), in contrast to the approximately 11% reduction in the density of asymmetric synapses (P = 0.20, Fig. 5C,D and Table 3). The density of asymmetric axospinous synapses was reduced by approximately 17% with age (P = 0.09 2-tailed Student's t-test; P = 0.05 1-tailed Student's t-test, justified by the significant loss of dendritic spines). The approximately 17% reduction in axospinous synapses is quite consistent with the approximately 24% reduction in spines across whole cells. The decrease in the numerical density of symmetric synapses with age was due to a significant decrease in the density of both symmetric axospinous (P = 0.002, ∼43% reduction) and axodendritic (P = 0.001, ∼53% reduction) synapses, which are lost at a similar rate. There was no difference in the synapse density between AI and AU subjects (asymmetric, P = 0.37 and symmetric, P = 0.99), and CII did not correlate with the density of either asymmetric or symmetric synapses (P = 0.17 and 0.16, respectively; not shown). There was no significant difference in the lengths of profiles of synaptic junctions in aged versus young V1 neurons, with the exception of symmetric axodendritic synapses, which were significantly longer in aged versus young neurons (Table 3; P = 0.04). These data on synapse profile lengths represent only general estimates of synapse size, but are consistent with the confocal data, in which no significant differences in the maximal widths of spine heads (a parameter also correlated with synapse size; Bourne and Harris 2008) in aged versus young neurons were observed.

Table 3

Synapse densities and profile lengths

Young Aged P-value
Numerical density, NV (synapses ×106/mm3
All synapses 598 ± 38 467 ± 18 0.002
Asymmetric 416 ± 36 369 ± 21 0.20
Axospinous 280 ± 25 231 ± 16 0.09
Axodendritic 100 ± 10 97 ± 10 0.85
Symmetric 186 ± 21 106 ± 6 0.0003
Axospinous 88 ± 12 50 ± 4 0.002
Axodendritic 88 ± 14 41 ± 4 0.001
Synapse profile length (µm)
Asymmetric (all) 0.34 ± 0.01 0.36 ± 0.01 0.08
Axospinous 0.33 ± 0.01 0.36 ± 0.01 0.06
Axodendritic 0.34 ± 0.02 0.35 ± 0.02 0.61
Symmetric (all) 0.29 ± 0.01 0.33 ± 0.02 0.06
Axospinous 0.28 ± 0.01 0.31 ± 0.01 0.15
Axodendritic 0.29 ± 0.02 0.35 ± 0.02 0.03
Young Aged P-value
Numerical density, NV (synapses ×106/mm3
All synapses 598 ± 38 467 ± 18 0.002
Asymmetric 416 ± 36 369 ± 21 0.20
Axospinous 280 ± 25 231 ± 16 0.09
Axodendritic 100 ± 10 97 ± 10 0.85
Symmetric 186 ± 21 106 ± 6 0.0003
Axospinous 88 ± 12 50 ± 4 0.002
Axodendritic 88 ± 14 41 ± 4 0.001
Synapse profile length (µm)
Asymmetric (all) 0.34 ± 0.01 0.36 ± 0.01 0.08
Axospinous 0.33 ± 0.01 0.36 ± 0.01 0.06
Axodendritic 0.34 ± 0.02 0.35 ± 0.02 0.61
Symmetric (all) 0.29 ± 0.01 0.33 ± 0.02 0.06
Axospinous 0.28 ± 0.01 0.31 ± 0.01 0.15
Axodendritic 0.29 ± 0.02 0.35 ± 0.02 0.03

P-values that are statistically significant (less than 0.05) are indicated in italics.

### In Vitro Electrophysiologic Properties of Layer 3 Pyramidal Cells in V1 Do Not Change with Age

In this study, a total of 30 young and 26 aged neurons were comprehensively electrophysiologically characterized. All the neurons exhibited regular spiking firing activity, although a subset of approximately 30% of the neurons also exhibited a phasic pattern of firing. The proportion of neurons showing regular versus phasic firing patterns did not differ between cells from young and aged subjects. Despite the significant alterations to the spines of these neurons and reduction in synapse densities, none of the 20 different electrophysiologic parameters examined differed substantially between young and aged neurons (Table 4). Thus, basic properties such as resting membrane potential, membrane time constant, and input resistance did not differ in the young versus aged layer 3 pyramidal cells. Single APs did not differ with regard to threshold, amplitude, or kinetics (Fig. 6A). Repetitive AP firing rates in response to sustained depolarizing current steps also did not differ (Fig. 6B,C), nor did the amplitude of the slow afterhyperpolarization (sAHP) current. Finally, with the exception of a longer time course (rise, half-width, and decay) of sEPSCs in aged neurons (P < 0.05), the mean frequency, amplitude, and kinetics of spontaneous EPSC and IPSC were not different in cells from young and aged subjects (Fig. 6D and Table 4). As with the morphologic findings, there was no difference in the electrophysiologic properties of neurons from the impaired and unimpaired aged subjects.

Table 4

Electrophysiologic properties

Young Aged P-value
Passive properties
Tau (ms) 28.4 ± 2.9 25.7 ± 3.0 0.52
Rn (MΩ) 285 ± 25.5 291 ± 27 0.87
Vr (mV) −65.8 ± 0.83 −64.3 ± 1.0 0.22
AP properties
Threshold (mV) −41.8 ± 0.74 −41.1 ± 0.7 0.52
Amp. (mV) 71.0 ± 1.6 73.2 ± 1.9 0.38
Dur. at 1/2 (ms) 1.44 ± 0.05 1.57 ± 0.06 0.11
Rise (ms) 0.91 ± 0.03 0.98 ± 0.04 0.15
Fall (ms) 2.12 ± 0.07 2.27 ± 0.10 0.21
Rheobase
Amp. (pA) 87.2 ± 11.5 89.8 ± 13.2 0.88
Medium AHP
Amp. (mV) −10.8 ± 0.66 −9.93 ± 0.37 0.31
Dur. (ms) 67.0 ± 14.3 52.6 ± 6.5 0.40
Slow AHP
Amp. (mV) −2.23 ± 0.49 −1.71 ± 0.28 0.37
Dur. (ms) 691 ± 23 659 ± 42 0.47
AP firing rates (Hz)
80-pA step 12.9 ± 1.2 11.5 ± 1.4 0.44
100-pA step 13.2 ± 1.1 14.0 ± 1.4 0.65
120-pA step 12.4 ± 1.4 14.9 ± 1.6 0.27
IH Amp. (pA)
−140 mV step 221 ± 33 216 ± 43 0.91
IsAHP Amp. (pA)
20-mV step 65.6 ± 8.5 85.4 ± 17.3 0.27
EPSCs
Frequency 1.51 ± 0.34 1.27 ± 0.17 0.48
Amplitude 6.43 ± 0.2 6.80 ± 0.3 0.34
Area 42.9 ± 3.0 50.9 ± 3.7 0.10
Rise 1.00 ± 0.06 1.26 ± 0.1 0.05
Half-width 4.38 ± 0.21 5.57 ± 0.44 0.03
Decay 4.22 ± 0.28 5.49 ± 0.5 0.05
IPSCs
Frequency 0.33 ± 0.08 0.20 ± 0.06 0.11
Amplitude 20.2 ± 1.1 19.6 ± 1.0 0.71
Area 308 ± 24 299 ± 25 0.81
Rise 2.78 ± 0.31 3.16 ± 0.30 0.39
Half-width 9.46 ± 0.53 10.5 ± 1.1 0.36
Decay 7.62 ± 0.60 7.63 ± 0.76 0.99
Young Aged P-value
Passive properties
Tau (ms) 28.4 ± 2.9 25.7 ± 3.0 0.52
Rn (MΩ) 285 ± 25.5 291 ± 27 0.87
Vr (mV) −65.8 ± 0.83 −64.3 ± 1.0 0.22
AP properties
Threshold (mV) −41.8 ± 0.74 −41.1 ± 0.7 0.52
Amp. (mV) 71.0 ± 1.6 73.2 ± 1.9 0.38
Dur. at 1/2 (ms) 1.44 ± 0.05 1.57 ± 0.06 0.11
Rise (ms) 0.91 ± 0.03 0.98 ± 0.04 0.15
Fall (ms) 2.12 ± 0.07 2.27 ± 0.10 0.21
Rheobase
Amp. (pA) 87.2 ± 11.5 89.8 ± 13.2 0.88
Medium AHP
Amp. (mV) −10.8 ± 0.66 −9.93 ± 0.37 0.31
Dur. (ms) 67.0 ± 14.3 52.6 ± 6.5 0.40
Slow AHP
Amp. (mV) −2.23 ± 0.49 −1.71 ± 0.28 0.37
Dur. (ms) 691 ± 23 659 ± 42 0.47
AP firing rates (Hz)
80-pA step 12.9 ± 1.2 11.5 ± 1.4 0.44
100-pA step 13.2 ± 1.1 14.0 ± 1.4 0.65
120-pA step 12.4 ± 1.4 14.9 ± 1.6 0.27
IH Amp. (pA)
−140 mV step 221 ± 33 216 ± 43 0.91
IsAHP Amp. (pA)
20-mV step 65.6 ± 8.5 85.4 ± 17.3 0.27
EPSCs
Frequency 1.51 ± 0.34 1.27 ± 0.17 0.48
Amplitude 6.43 ± 0.2 6.80 ± 0.3 0.34
Area 42.9 ± 3.0 50.9 ± 3.7 0.10
Rise 1.00 ± 0.06 1.26 ± 0.1 0.05
Half-width 4.38 ± 0.21 5.57 ± 0.44 0.03
Decay 4.22 ± 0.28 5.49 ± 0.5 0.05
IPSCs
Frequency 0.33 ± 0.08 0.20 ± 0.06 0.11
Amplitude 20.2 ± 1.1 19.6 ± 1.0 0.71
Area 308 ± 24 299 ± 25 0.81
Rise 2.78 ± 0.31 3.16 ± 0.30 0.39
Half-width 9.46 ± 0.53 10.5 ± 1.1 0.36
Decay 7.62 ± 0.60 7.63 ± 0.76 0.99
Figure 6.

AP and synaptic properties do not differ with age. (A) Superimposed representative APs from young (black) and aged (gray) cells. (B) Voltage–frequency plot showing no change in AP firing rates with age. (C) Voltage responses to depolarizing current ramps of representative young (black) and aged (gray) neurons (top traces) and to +80 pA 2 s step stimuli. (D) Top, averaged EPSCs from representative young and aged neurons. Bottom, averaged IPSCs from representative young and aged neurons. Scale bars: A = 20 mV/5 ms; C, top = 20 mV/500 ms; C, bottom = 20 mV/2 s; D, top = 2 pA/5 ms; D, bottom, 2 pA/10 ms.

Figure 6.

AP and synaptic properties do not differ with age. (A) Superimposed representative APs from young (black) and aged (gray) cells. (B) Voltage–frequency plot showing no change in AP firing rates with age. (C) Voltage responses to depolarizing current ramps of representative young (black) and aged (gray) neurons (top traces) and to +80 pA 2 s step stimuli. (D) Top, averaged EPSCs from representative young and aged neurons. Bottom, averaged IPSCs from representative young and aged neurons. Scale bars: A = 20 mV/5 ms; C, top = 20 mV/500 ms; C, bottom = 20 mV/2 s; D, top = 2 pA/5 ms; D, bottom, 2 pA/10 ms.

### Outward Voltage Attenuation Is Reduced in Modeled Aged Versus Young Apical Dendrites

Computational modeling was used to evaluate whether the morphologic changes observed with aging in V1 neurons might affect the attenuation of electrical signals through the dendrites. Electrotonic analyses were performed on 8 young and 7 aged neurons, using passive parameters fit to V1 physiologic data in Amatrudo et al. (2012). Figure 7A shows the morphoelectrotonic transforms of outward and inward electrical signals (Lout and Lin at DC, 0 Hz) for the neurons pictured in Figure 1A. These transforms scale each anatomical branch according to how much voltage attenuates, either propagating away from the soma (outward) or propagating in toward it (inward), allowing visual comparison of voltage attenuation in the 2 neurons. Figure 7B,C shows the mean attenuation lengths $L¯out$ and Lin of the young and aged neurons, respectively, as input frequency varies from 0 to 500 Hz, separated into apical and basal arbors. Our previous studies (Kabaso et al. 2009; Amatrudo et al. 2012) have shown that $L¯out$ and Lin correlate well with functional measures such as passive AP backpropagation and the amplitudes of EPSCs simulated across the dendritic arbors.

Figure 7.

Electrotonic analyses of young and aged neurons. (A) Morphoelectrotonic transforms of the representative young and aged neurons shown in Figure 1. Scale bars represent 1/10th of an attenuation unit in the outward transforms (left), and one attenuation unit in the inward transforms (right). Transforms of apical arbors are shown in red, and basal arbors in black. (B) Mean outward attenuation, $L¯out$, versus input frequency (Hz) of the young and aged V1 neurons, in black and gray, respectively. (C) Analogous to (B), for mean inward attenuation, $L¯in$, versus input frequency. *P < 0.05.

Figure 7.

Electrotonic analyses of young and aged neurons. (A) Morphoelectrotonic transforms of the representative young and aged neurons shown in Figure 1. Scale bars represent 1/10th of an attenuation unit in the outward transforms (left), and one attenuation unit in the inward transforms (right). Transforms of apical arbors are shown in red, and basal arbors in black. (B) Mean outward attenuation, $L¯out$, versus input frequency (Hz) of the young and aged V1 neurons, in black and gray, respectively. (C) Analogous to (B), for mean inward attenuation, $L¯in$, versus input frequency. *P < 0.05.

Mean outward attenuation was significantly lower in the apical arbors of aged versus young neurons at every input frequency (Fig. 7B, P ≤ 0.05 for each). Even when spines were excluded from the calculations, outward attenuation in the apical arbors was lower in the aged neurons for inputs at 100 and 200 Hz (P = 0.048 for both). Mean outward attenuation in the basal arbors did not differ in aged versus young neurons (Fig. 7B, P ≥ 0.20 for all frequencies), either with or without spines. Similarly, there was no difference in mean inward attenuation in aged versus young neurons for either the apical or basal arbors (Fig. 7C, P ≥ 0.20 for all frequencies). This is consistent with the empirical finding that EPSC and IPSC amplitudes do not change with aging in V1 L3 pyramidal neurons. While empirically assessed electrophysiologic properties did not differ in young versus aged neurons (Table 4), these computational electrotonic analyses predict that a signal propagating from the soma out into the apical dendrites attenuates less in the aged V1 neurons than in the young neurons.

## Discussion

This study represents the first comprehensive investigation of the effects of normal aging on detailed structural and electrophysiologic properties of individual layer 3 pyramidal cells of the macaque monkey V1. A number of significant structural alterations occur in aged neurons, including regression of the distal apical dendritic arbor, a loss of approximately 24% of dendritic spines, with a specific reduction in mushroom spines in the apical arbor and thin spines in the basilar arbor, and a reduced numerical density of symmetric synapses in the neuropil of layers 2–3. The electrophysiologic properties of aged and young neurons do not differ, although modeling predicts that backpropagating APs (not measured empirically here) would attenuate less in aged than in young neurons. The age-related alterations to neurons and neuropil do not appear to be related to cognitive impairment, as data from cognitively impaired and cognitively spared aged monkeys does not differ.

### Apical Dendritic Regression in Aged Neurons

The total dendritic length of aged neurons does not differ from that of young neurons, but there is a significant reduction in both the number of dendritic intersections and dendritic length in the distal third of the apical tree in aged neurons. The decrease in dendritic length and complexity limited to the distal portion of the apical tree is consistent with our earlier report that the thickness of layer 1 in the monkey V1 decreases significantly with age, and contains fewer dendritic branch and spine profiles (Peters et al. 2001). The apical tufts of layer 3 pyramidal cells contribute substantially to the neuropil of layer 1, making it likely that the thinning of layer 1 results in large part from the regression of the distal apical region of these pyramidal cells. Interestingly, the specific regression of the distal apical dendritic branches mirrors what has been reported for layer 3 neurons in the aging monkey PFC (Kabaso et al. 2009), indicating that this part of the dendritic arbor is especially vulnerable during normal aging across diverse neocortical areas. Given that the distal dendritic arbor represents a significant proportion of the total dendritic surface area of pyramidal neurons and receives important thalamic, cortical feedback, and neuromodulatory inputs (reviewed in Jones 2001; Callaway 2004; Spruston 2008), regression of this dendritic substrate may have particularly important consequences with regard to signal integration by pyramidal neurons.

### Loss of Dendritic Spines in Aged Neurons

In aged neurons, there is a significant approximately 24% decrease in the total density of dendritic spines. This loss of spines is somewhat smaller in magnitude than the 33–47% loss reported for aged PFC neurons (Duan et al. 2003; Dumitriu et al. 2010), which have much higher numbers and densities of spines than do V1 neurons (Amatrudo et al. 2012). Interestingly, the loss of spines is subtype-specific, and differs between the apical and basal arbors. Thus, mushroom spines are selectively lost from the apical arbor, while thin, stubby, and filopodia densities are unchanged, and thin spines are selectively lost from the basal arbor, while mushroom, stubby, and filopodia densities are not altered. This latter finding is consistent with the specific loss of thin spines in monkey PFC pyramidal neurons previously reported (Duan et al. 2003; Dumitriu et al. 2010). It has been hypothesized that the specific loss of thin spines, which are considered highly plastic and required for learning new information (Kasai et al. 2003, 2010; Bourne and Harris 2007), may underlie impairment in acquisition of cognitive tasks in aging primates (Dumitriu et al. 2010; Morrison and Baxter 2012). On the other hand, in V1 neurons, the specific loss of mushroom spines, which are the largest and most stable spine subtype, could plausibly contribute to impaired visual selectivity. A specific loss of mushroom spines has previously been reported in dentate gyrus granule cells of aging primates (Hara et al. 2012), and this loss was associated with impairment on the DNMS task accuracy. Thus, it appears that, with normal aging, loss of dendritic spines is ubiquitous across multiple brain areas and neuron types, but that different spine subtypes are differentially vulnerable depending on the brain area assessed. Here, we provide evidence that specific spine subtype loss can also differ between apical and basal dendritic arbors of the same individual V1 neurons. Interestingly, a similar heterogeneity in age-related spine loss from apical versus basal dendritic arbors has previously been reported by Benavides-Piccione et al. (2013) in human cingulate cortex pyramidal cells, which exhibit a loss of small, short spines from the basal arbor and of long spines from the apical arbor. Apical and basal dendrites arborize in distinct layers that have distinct connectional microenvironments and functional properties in V1 (reviewed in Hirsch and Martinez 2006). Thus, this finding has implications for differences in the effects of normal aging on signal processing and integration between the apical and basal arbors, which play distinctive roles in the overall excitability of layer 3 pyramidal neurons in V1.

### Reduction in the Numerical Density of Synapses in Aged V1 Neuropil

Electron microscopic analyses demonstrate that there is an overall reduction of synapses in the neuropil of layers 2–3 of V1 in aged versus young monkeys, although the population of asymmetric “excitatory” synapses is relatively less affected than that of symmetric “inhibitory” synapses (reductions of ∼11% and ∼43%, respectively). While the numerical density of total asymmetric synapses does not differ significantly, the specific population of asymmetric axospinous synapses is reduced by approximately 17% in aged V1. This is consistent with, but not identical to, the approximately 24% loss of dendritic spines on layer 3 pyramidal neurons. This modest discrepancy is likely due to sampling differences between individual layer 3 pyramidal neurons versus the neuropil as a whole, which contains dendrites from additional neuronal populations (e.g., layer 5 pyramidal cells), and to the fact that neuropil was assessed in layers 2–3 only, while spines were assessed across layers 1, 2, and 3. In addition, although the approaches used here for single neuron imaging and data rendering are robust and have been used previously to express age- or disease-related differences in spine density (for review, see Luebke, Weaver, et al. 2010; Dickstein et al. 2013), the difference in resolution between confocal microscopy and EM could also lead to a population of the smallest spines being visible and included at the EM level, but not present in the confocal dataset (given the ∼0.2 µm theoretical resolution and ∼0.4 µm axial resolution of laser-scanning confocal microscopy). Questions thus arise as to which population of dendrites in the neuropil exhibits loss of spines and synapses with age, and whether there are undetected subtle age-related changes in synapse size and shape that would have implications for the distribution of synapses. Further studies using labeling of specific populations of neurons and 3D reconstructions from serial EM sections will be needed to address these issues. Nevertheless, the result for asymmetric synapses in aged V1 differs from the 30% age-related reduction of asymmetric synapses shown in our previous study of the dorsolateral PFC (Peters, Sethares, et al. 2008), consistent with the difference in the magnitude of spine loss seen in the 2 cortical areas.

The most prominent age-related structural change in V1 is that of a dramatic 43% reduction in the numerical density of symmetric synapses on spines and dendrites in the neuropil of layers 2–3. The loss of symmetric synapses is consistent with studies in human V1, wherein distal dendritic- and spine-targeting calbindin-expressing inhibitory neurons are reduced in number with age (Bu et al. 2003). In contrast, the density of parvalbumin-expressing perisomatic-targeting inhibitory neurons is unchanged with age (Bu et al. 2003; reviewed in Lehmann et al. 2012). While it was not possible to specify the exact dendritic domain of the synapses counted, the fields sampled here excluded somatic and proximal dendritic domains. Differential effects of aging on somatic versus dendritic inhibition may explain why the 43% reduction in inhibitory synapses with age did not translate into reduced inhibitory synaptic current frequency or amplitude in the in vitro recording studies. It is possible that the magnitude of loss of inhibitory synapses in V1 neurons is, indeed, functionally more compartmentalized to distal dendritic segments, while whole-cell, patch-clamp recording from the somata of neurons in in vitro slices is only capable of assessing relatively proximal synaptic inputs due to space clamp limitations. Interestingly, age-related visual deficits and single-unit activity abnormalities assessed in monkeys in vivo are ameliorated by application of the GABAA receptor agonist muscimol (Leventhal et al. 2003). The present findings of significant loss of symmetric synapses in V1 with age may well provide the structural basis for this important observation.

In the rhesus monkey dorsolateral PFC, a similar but less pronounced reduction in the density of axodendritic and axospinous inhibitory synapses in layers 2–3 was found with age (Peters, Sethares, et al. 2008). However, these pyramidal neurons show an increase in the frequency of spontaneous IPSCs with age (Luebke et al. 2004), which suggests a compensatory increase of activity in the remaining inhibitory axon terminals. Indeed, a follow-up study showed that the density of axosomatic symmetric synapses was unchanged with age, but the axon terminals that form these synapses showed an increase in mitochondrial size and number of vesicles (Soghomonian et al. 2010). A similar compensatory mechanism may occur in layers 2–3 of V1 that increases the synaptic efficacy of the remaining inhibitory terminals, especially axosomatic symmetric synapses, to maintain the frequency and amplitude of IPSCs with age.

Although further confirmatory studies are needed, our data are consistent with previously shown downregulation of dendritic inhibition and resilience of somatic inhibition with age in mammalian V1 (Bu et al. 2003; Atallah et al. 2012; Mao et al. 2012). Interestingly, dendritic inhibition is thought to be key for tuning and selectivity in cortical processing (Wang et al. 2004; reviewed in Lehmann et al. 2012). In mouse V1, genetically induced loss of dendrite-targeting inhibitory neurons has been shown to lead to impairments in orientation tuning (Mao et al. 2012), while activation of soma-targeting parvalbumin-expressing inhibitory neurons has no effect on orientation selectivity (Atallah et al. 2012).

### Preservation of Physiologic Properties of Aged Neurons

In contrast to the fairly robust structural changes observed, the fundamental electrophysiologic properties of layer 3 pyramidal cells did not change with age. This lack of an effect is in marked contrast to the changes seen in aged layer 3 PFC neurons in vitro. These include significantly increased input resistance and AP firing rates (Chang et al. 2005), significantly increased sAHP current amplitude (Luebke and Amatrudo 2010), and reduced synaptic excitation and enhanced synaptic inhibition (Luebke et al. 2004). This difference in findings is interesting given that neurons in both brain areas undergo significant structural changes, albeit to different degrees. Compared with those in V1, layer 3 pyramidal neurons in PFC possess a much higher density of dendritic spines (for review, see Elston 2003, 2007), and more than a third of the spines on PFC neurons are lost during normal aging, while in V1 a more modest 24% of spines are lost. In PFC neurons, spontaneous excitatory synaptic events are significantly reduced in frequency, while the frequency of these events is unchanged in V1 neurons. It is possible that the magnitude of loss of spines and synapses in V1 neurons is, indeed, functionally important, but that our methodology (whole-cell, patch-clamp recording from the somata of neurons in in vitro slices) has insufficient spatial resolution to demonstrate the functional consequences of this loss. In summary, monkey PFC neurons appear to be more vulnerable than V1 neurons to the effects of aging, particularly with regard to physiologic changes. This heterogeneity in the effects of age on neurons in the 2 brain areas is consistent with the increasingly detailed evidence for significant differences in the normative properties of pyramidal neurons in V1 and PFC both in adulthood (Jacobs and Scheibel 2002; Elston 2003, 2007; Amatrudo et al. 2012) and during development (Jacobs et al. 2001; Bianchi et al. 2012, 2013; Oga et al. 2013).

Despite the similarity of electrophysiologic properties in young and aged neurons in V1, the modeling results predict that dendritic signal integration is altered with aging in these neurons. Unless the distribution of active ionic conductances across the dendrites is altered in aging, APs likely backpropagate further into the apical dendrites of aged V1 neurons than they do in young ones. This could have implications for synaptic plasticity and the generation of N-methyl-D-aspartate (NMDA)-dependent plateau potentials in the apical tuft (for review, see Waters et al. 2005; Major et al. 2013).

### Functional Implications

Previous studies have demonstrated that cortical pyramidal neurons from AI and AU animals can possess significantly different structural and functional properties (Moyer et al. 2000; Tombaugh et al. 2005; Matthews et al. 2009; Luebke and Amatrudo 2010; for review, see Morrison and Baxter 2012). In the present study, we similarly compared data from impaired and unimpaired aged subjects and found no difference in any of the variables between these groups. This was not entirely unexpected, given that V1 is a primary sensory area critical for early visual processing but not directly involved in higher-order cognitive functions per se.

In rhesus monkey V1, Leventhal et al. (2003) have demonstrated a degradation of orientation and direction selectivity with age, correlated with an increase in spontaneous activity and responses to all stimuli in V1 neurons in vivo (Schmolesky et al. 2000; Fu et al. 2013). Moreover, these age-related visual deficits and single-unit activity abnormalities assessed in vivo could be improved by application of the GABAA receptor agonist muscimol. These functional findings are consistent with the structural changes shown here. The regression of distal apical dendrites suggests a loss of feedback and neuromodulatory inputs, which are important for spatial tuning of pyramidal neuronal output (Spruston 2008). In addition, the loss of inhibitory synapses with age shown here is consistent with the functional reduction in inhibitory tone with age shown in vivo (Schmolesky et al. 2000).

In conclusion, while the structural changes we observed did not impact the physiologic properties of single neurons assessed in vitro, the additive effects of structural alterations has significant implications for visual processing by neurons embedded in a local V1 network in vivo. Deficits in visual function with age are thus likely due to the combined effects of diverse age-related structural changes, including spine and synapse loss, dendritic and axonal regression, and myelin dystrophy.

## Funding

This work was supported by the National Institutes of Health (grants P01 AG00001, R01 AG025062, R01 AG035071, R01 MH071818, and R01 DC05669).

## Notes

We thank Claire Sethares for expert assistance with electron microscopy. Conflict of Interest: None declared.

## References

A
Spear
PD
.
1993
.
Effects of aging on the size, density, and number of rhesus monkey lateral geniculate neurons
.
J Comp Neurol
.
334
:
631
643
.
Amatrudo
JM
Weaver
CM
Crimins
JL
Hof
PR
Rosene
DL
Luebke
JI
.
2012
.
Influence of highly distinctive structural properties on the excitability of pyramidal neurons in monkey visual and prefrontal cortices
.
J Neurosci
.
32
(40)
:
13644
13660
.
Atallah
BV
Bruns
W
Carandini
M
Scanziani
M
.
2012
.
Parvalbumin-expressing interneurons linearly transform cortical responses to visual stimuli
.
Neuron
.
73
(1)
:
159
170
.
Barnes
CS
De L'Aune
W
Schuchard
RA
.
2011
.
A test of face discrimination ability in aging and vision loss
.
Optom Vis Sci
.
88
:
188
199
.
Benavides-Piccione
R
Fernaud-Espinosa
I
Robles
V
Yuste
R
Defelipe
J
.
2013
.
Age-based comparison of human dendritic spine structure using complete three-dimensional reconstructions
.
Cereb Cortex
.
23
(8)
:
1798
1810
.
Betts
LR
Sekuler
AB
Bennett
PJ
.
2007
.
The effects of aging on orientation discrimination
.
Vision Res
.
47
:
1769
1780
.
Bianchi
S
Stimpson
CD
Bauernfeind
AL
Schapiro
SJ
Baze
WB
McArthur
MJ
Bronson
E
Hopkins
WD
Semendeferi
K
Jacobs
B
et al
2012
.
Dendritic morphology of pyramidal neurons in the chimpanzee neocortex: regional specializations and comparison to humans
.
Cereb Cortex
.
10
:
2429
2436
.
Bianchi
S
Stimpson
CD
Duka
T
Larsen
MD
Janssen
WG
Collins
Z
Bauernfeind
AL
Schapiro
SJ
Baze
WB
McArthur
MJ
et al
2013
.
Synaptogenesis and development of pyramidal neuron dendritic morphology in the chimpanzee neocortex resembles humans
.
.
110
(Suppl 2)
:
10395
10401
.
Bourne
J
Harris
KM
.
2007
.
Do thin spines learn to be mushroom spines that remember?
Curr Opin Neurobiol
.
17
:
381
386
.
Bourne
JN
Harris
KM
.
2008
.
Balancing structure and function at hippocampal dendritic spines
.
Annu Rev Neurosci
.
31
:
47
67
.
Brown
TH
A
Mainen
ZF
Claiborne
BJ
.
1992
.
Hebbian computations in hippocampal dendrites and spines
. In:
McKenna
T
Davis
J
Zornetzer
SF
, editors.
Single neuron computation
.
San Diego
:
, pp.
81
116
.
Bu
J
Sathyendra
V
Nagykery
N
Geula
C
.
2003
.
Age-related changes in calbindin-D28k, calretinin, and parvalbumin immunoreactive neurons in the human cerebral cortex
.
Exp Neurol
.
182
(1)
:
220
231
.
Callaway
EM
.
2004
.
Feedforward, feedback and inhibitory connections in primate visual cortex
.
Neural Netw
.
17
(5–6)
:
625
632
.
Calverley
RK
Bedi
KS
Jones
DG
.
1988
.
Estimation of the numerical density of synapses in rat neocortex. Comparison of the ‘disector’ with an ‘unfolding’ method
.
J Neurosci Methods
.
23
(3)
:
195
205
.
Carnevale
NT
Hines
ML
.
2006
.
The NEURON Book
.
Cambridge, UK
:
Cambridge University Press
.
Carnevale
NT
Tsai
KY
Claiborne
BJ
Brown
TH
.
1997
.
Comparative electrotonic analysis of three classes of rat hippocampal neurons
.
J Neurophysiol
.
78
:
703
720
.
Chang
Y
Rosene
DL
Killiany
RJ
Mangiamele
LA
Luebke
JL
.
2005
.
Increased action potential firing rates of layer 2–3 pyramidal cells in the prefrontal cortex are significantly related to cognitive performance in aged monkeys
.
Cereb Cortex
.
15
:
409
418
.
Colonnier
M
Beaulieu
C
.
1985
.
An empirical assessment of stereological formulae applied to the counting of synaptic disks in the cerebral cortex
.
J Comp Neurol
.
231
(2)
:
175
179
.
DeFelipe
J
Marco
P
Busturia
I
Merchán-Pérez
A
.
1999
.
Estimation of the number of synapses in the cerebral cortex: methodological considerations
.
Cereb Cortex
.
9
(7)
:
722
732
.
Delahunt
PB
Hardy
JL
Werner
JS
.
2008
.
The effect of senescence on orientation discrimination and mechanism tuning
.
J Vis
.
8
:
5.1
5.9
.
Dickstein
DL
Weaver
CM
Luebke
JI
Hof
PR
.
2013
.
Dendritic spine changes associated with normal aging
.
Neuroscience
.
251
:
21
32
.
Duan
H
Wearne
SL
Rocher
AB
Macedo
A
Morrison
JH
Hof
PR
.
2003
.
Age-related morphologic alterations in dendrites and spine densities of corticocortically projecting neurons in macaque monkeys
.
Cereb Cortex
.
13
:
950
961
.
Dumitriu
D
Hao
J
Hara
Y
Kaufmann
J
Janssen
WG
Lou
W
Rapp
PR
Morrison
JH
.
2010
.
Selective changes in thin spine density and morphology in monkey prefrontal cortex correlate with aging-related cognitive impairment
.
J Neurosci
.
30
(22)
:
7507
7515
.
Elliott
DB
Whitaker
D
MacVeigh
D
.
1990
.
Neural contribution to spatiotemporal contrast sensitivity decline in healthy eyes
.
Vision Res
.
30
:
541
547
.
Elliott
SL
Choi
SS
Doble
N
Hardy
JL
Evans
JW
Werner
JS
.
2009
.
Role of high-order aberrations in senescent changes in spatial vision
.
J Vis
.
9
:
1
16
.
Elston
GN
.
2003
.
Cortex, cognition and the cell: new insights into the pyramidal neuron and prefrontal function
.
Cereb Cortex
.
13
:
1124
1138
.
Elston
GN
.
2007
.
Specializations in pyramidal cell structure during primate evolution
. In:
Kaas
JH
Preuss
TM
, editors.
Evolution of nervous systems
.
Oxford
:
, pp.
191
242
.
Fiala
JC
.
2005
.
Reconstruct: a free editor for serial section microscopy
.
J Microsc
.
218
:
52
61
.
Fu
Y
Yu
S
Ma
Y
Wang
Y
Zhou
Y
.
2013
.
Functional degradation of the primary visual cortex during early senescence in rhesus monkeys
.
Cereb Cortex
.
23
:
2923
2931
.
Giannaris
EL
Rosene
DL
.
2012
.
A stereological study of the numbers of neurons and glia in the primary visual cortex across the lifespan of male and female rhesus monkeys
.
J Comp Neurol
.
520
(15)
:
3492
3508
.
Hao
J
Rapp
PR
Leffler
AE
Leffler
SR
Janssen
WG
Lou
W
McKay
H
Roberts
JA
Wearne
SL
Hof
PR
et al
2006
.
Estrogen alters spine number and morphology in prefrontal cortex of aged female rhesus monkeys
.
J Neurosci
.
26
:
2571
2578
.
Hara
Y
Punsoni
M
Yuk
F
Park
CS
Janssen
WG
Rapp
PR
Morrison
JH
.
2012
.
Synaptic distributions of GluA2 and PKMζ in the monkey dentate gyrus and their relationships with aging and memory
.
J Neurosci
.
32
(21)
:
7336
7344
.
Herndon
JG
Moss
MB
Rosene
DL
Killiany
RJ
.
1997
.
Patterns of cognitive decline in aged rhesus monkeys
.
Behav Brain Res
.
87
:
25
35
.
Hirsch
JA
Martinez
LM
.
2006
.
Laminar processing in the visual cortical column
.
Curr Opin Neurobiol
.
16
(4)
:
377
384
.
Hof
PR
Nimchinsky
EA
Young
WG
Morrison
JH
.
2000
.
Numbers of Meynert and layer IVB cells in area V1: a stereologic analysis in young and aged macaque monkeys
.
J Comp Neurol
.
420
(1)
:
113
126
.
Jacobs
B
Schall
M
Prather
M
Kapler
E
Driscoll
L
Baca
S
Jacobs
J
Ford
K
Wainwright
M
Treml
M
.
2001
.
Regional dendritic and spine variation in human cerebral cortex: a quantitative study
.
Cereb Cortex
.
11
:
558
571
.
Jacobs
B
Scheibel
AB
.
2002
.
Regional dendritic variation in primate cortical pyramidal cells
. In:
Schüz
A
Miller
R
, editors.
Cortical areas: unity and diversity
.
London
:
Taylor and Francis
, pp.
111
131
.
Jones
EG
.
2001
.
The thalamic matrix and thalamocortical synchrony
.
Trends Neurosci
.
24
(10)
:
595
601
.
Kabaso
D
Coskren
PJ
Henry
BI
Hof
PR
Wearne
SL
.
2009
.
The electrotonic structure of pyramidal neurons contributing to prefrontal cortical circuits in macaque monkeys is significantly altered in aging
.
Cereb Cortex
.
19
:
2248
2268
.
Kasai
H
Hayama
T
Ishikawa
M
Watanabe
S
Yagishita
S
Noguchi
J
.
2010
.
Learning rules and persistence of dendritic spines
.
Eur J Neurosci
.
32
:
241
249
.
Kasai
H
Matsuzaki
M
Noguchi
J
Yasumatsu
N
Nakahara
H
.
2003
.
Structure-stability-function relationships of dendritic spines
.
Trends Neurosci
.
26
:
360
368
.
Kim
CB
Tom
BW
Spear
PD
.
1996
.
Effects of aging on the densities, numbers, and sizes of retinal ganglion cells in rhesus monkey
.
Neurobiol Aging
.
17
:
431
438
.
Lehmann
K
Steinecke
A
Bolz
J
.
forthcoming 2012
.
GABA through the ages: regulation of cortical function and plasticity by inhibitory interneurons
.
Neural Plast
.
Leventhal
AG
Wang
Y
Pu
M
Zhou
Y
Ma
Y
.
2003
.
GABA and its agonists improved visual cortical function in senescent monkeys
.
Science
.
300
:
812
815
.
Luebke
JI
Chang
Y
Moore
TL
Rosene
DL
.
2004
.
Normal aging results in decreased synaptic excitation and increased synaptic inhibition of layer 2–3 pyramidal cells in the monkey prefrontal cortex
.
Neuroscience
.
125
:
277
288
.
Luebke
JI
Weaver
CM
Rocher
AB
Rodriguez
A
Crimins
JL
Dickstein
DL
Wearne
SL
Hof
PR
.
2010
.
Dendritic vulnerability in neurodegenerative disease: insights from analyses of cortical pyramidal neurons in transgenic mouse models
.
Brain Struct Funct
.
214
:
181
199
.
Luebke
JL
Amatrudo
JM
.
2010
.
Age-related increase of sI(AHP) in prefrontal pyramidal cells of monkeys: relationship to cognition
.
Neurobiol Aging
.
33
:
1085
1095
.
Luebke
JL
Barbas
H
Peters
A
.
2010
.
Effects of normal aging on prefrontal area 46 in the rhesus monkey
.
Brain Res Rev
.
62
:
212
232
.
Major
G
Larkum
ME
Schiller
J
.
2013
.
Active properties of neocortical pyramidal neuron dendrites
.
Annu Rev Neurosci
.
36
:
1
24
.
Mao
R
Schummers
J
Knoblich
U
Lacey
CJ
Van Wart
A
Cobos
I
Kim
C
Huguenard
JR
Rubenstein
JL
Sur
M
.
2012
.
Influence of a subtype of inhibitory interneuron on stimulus-specific responses in visual cortex
.
Cereb Cortex
.
22
(3)
:
493
508
.
Matthews
EA
Linardakis
JM
Disterhoft
JF
.
2009
.
The fast and slow afterhyperpolarizations are differentially modulated in hippocampal neurons by aging and learning
.
J Neurosci
.
29
(15)
:
4750
4755
.
Moore
TL
Killiany
RJ
Herndon
JG
Rosene
DL
Moss
MB
.
2006
.
Executive system dysfunction occurs as early as middle-age in the rhesus monkey
.
Neurobiol Aging
.
27
:
1484
1493
.
Moore
TL
Killiany
RJ
Herndon
JG
Rosene
DL
Moss
MB
.
2003
.
Impairment in abstraction and set shifting in aged rhesus monkeys
.
Neurobiol Aging
.
24
:
125
134
.
Moore
TL
Schettler
SP
Killiany
RJ
Herndon
JG
Luebke
JI
Moss
MB
Rosene
DL
.
2005
.
Cognitive impairment in aged rhesus monkeys associated with monoamine receptors in the prefrontal cortex
.
Behav Brain Res
.
160
:
208
221
.
Morrison
JH
Baxter
MG
.
2012
.
The ageing cortical synapse: hallmarks and implications for cognitive decline
.
Nat Rev Neurosci
.
13
(4)
:
240
250
.
Moss
MB
Killiany
RJ
Lai
ZC
Rosene
DL
Herndon
JG
.
1997
.
Recognition memory span in rhesus monkeys of advanced age
.
Neurobiol Aging
.
18
(1)
:
13
19
.
Moss
MB
Rosene
DL
Peters
A
.
1988
.
Effects of aging on visual recognition memory in the rhesus monkey
.
Neurobiol Aging
.
9
(5–6)
:
495
502
.
Moyer
JR
Power
JM
Thompson
LT
Disterhoft
JF
.
2000
.
Increased excitability of aged rabbit CA1 neurons after trace eyeblink conditioning
.
J Neurosci
.
20
(14)
:
5476
5482
.
Nielsen
K
Peters
A
.
2000
.
The effects of aging on the frequency of nerve fibers in rhesus monkey striate cortex
.
Neurobiol Aging
.
21
(5)
:
621
628
.
Norman
JF
Burton
CL
Best
LA
.
2010
.
Modulatory effects of binocular disparity and aging upon the perception of speed
.
Vision Res
.
50
:
65
71
.
Norman
JF
Ross
HE
Hawkes
LM
Long
JR
.
2003
.
Aging and the perception of speed
.
Perception
.
32
:
85
96
.
Oga
T
Aoi
H
Sasaki
T
Fujita
I
Ichinohe
N
.
2013
.
Postnatal development of layer III pyramidal cells in the primary visual, inferior temporal, and prefrontal cortices of the marmoset
.
Front Neural Circuits
.
7
:
31
46
.
Owsley
C
.
2011
.
Aging and vision
.
Vision Res
.
51
:
1610
1622
.
Owsley
C
Sekuler
R
Boldt
C
.
1981
.
Aging and low-contrast vision: face perception
.
Invest Ophthalmol Vis Sci
.
21
(2)
:
362
365
.
Peters
A
Kemper
T
.
2012
.
A review of the structural alterations in the cerebral hemispheres of the aging rhesus monkey
.
Neurobiol Aging
.
33
(10)
:
2357
2372
.
Peters
A
Moss
MB
Sethares
C
.
2000
.
Effects of aging on myelinated nerve fibers in monkey primary visual cortex
.
J Comp Neurol
.
419
(3)
:
364
376
.
Peters
A
Nigro
NJ
McNally
KJ
.
1997
.
A further evaluation of the effect of age on striate cortex of the rhesus monkey
.
Neurobiol Aging
.
18
(1)
:
29
36
.
Peters
A
Palay
SL
Webster
HD
.
1991
The fine structure of the nervous system. Neurons and their supporting cells
.
New York
:
Oxford University Press
.
Peters
A
Sethares
C
.
2002
.
The effects of age on the cells in layer 1 of primate cerebral cortex
.
Cereb Cortex
.
12
(1)
:
27
36
.
Peters
A
Sethares
C
.
2003
.
Is there remyelination during aging of the primate central nervous system?
J Comp Neurol
.
460
(2)
:
238
254
.
Peters
A
Sethares
C
Killiany
RJ
.
2001
.
Effects of age on the thickness of myelin sheaths in monkey primary visual cortex
.
J Comp Neurol
.
435
(2)
:
241
248
.
Peters
A
Sethares
C
Luebke
JI
.
2008
.
Synapses are lost during aging in the primate prefrontal cortex
.
Neuroscience
.
152
(4)
:
970
981
.
Peters
A
Verderosa
A
Sethares
C
.
2008
.
The neuroglial population in the primary visual cortex of the aging rhesus monkey
.
Glia
.
56
(11)
:
1151
1161
.
Rodriguez
A
Ehlenberger
DB
Dickstein
DL
Hof
PR
Wearne
SL
.
2008
.
Automated three-dimensional detection and shape classification of dendritic spines from fluorescence microscopy images
.
PLoS One
.
3
:
e1997
.
Rodriguez
A
Ehlenberger
DB
Hof
PR
Wearne
SL
.
2006
.
Rayburst sampling, an algorithm for automated three-dimensional shape analysis from laser scanning microscopy images
.
Nat Protoc.

1
:
2152
2161
.
Rodriguez
A
Ehlenberger
DB
Kelliher
K
Einstein
M
Henderson
SC
Morrison
JH
Hof
PR
Wearne
SL
.
2003
.
Automated reconstruction of three-dimensional neuronal morphology from laser scanning microscopy images
.
Methods
.
30
:
94
105
.
Sandell
JH
Peters
A
.
2001
.
Effects of age on nerve fibers in the rhesus monkey optic nerve
.
J Comp Neurol
.
429
(4)
:
541
553
.
Schmolesky
MT
Wang
Y
Pu
M
Leventhal
AG
.
2000
.
Degradation of stimulus selectivity of visual cortical cells in senescent rhesus monkeys
.
Nat Neurosci
.
3
:
384
390
.
Scorcioni
R
Polavaram
S
Ascoli
GA
.
2008
.
L-Measure: a web-accessible tool for the analysis, comparison and search of digital reconstructions of neuronal morphologies
.
Nat Protoc
.
3
:
866
876
.
Sholl
DA
.
1953
.
Dendritic organization in the neurons of the visual and motor cortices of the cat
.
J Anat
.
87
:
387
406
.
Soghomonian
JJ
Sethares
C
Peters
A
.
2010
.
Effects of age on axon terminals forming axosomatic and axodendritic inhibitory synapses in prefrontal cortex
.
Neuroscience
.
168
(1)
:
74
81
.
Spear
PD
.
1993
.
Neural bases of visual deficits during aging
.
Vision Res
.
33
:
2589
2609
.
Spear
PD
Moore
RJ
Kim
CB
Xue
JT
Tumosa
N
.
1994
.
Effects of aging on the primate visual system: spatial and temporal processing by lateral geniculate neurons in young adult and old rhesus monkeys
.
J Neurophysiol
.
72
:
402
420
.
Spruston
N
.
2008
.
Pyramidal neurons: dendritic structure and synaptic integration
.
Nat Rev Neurosci
.
9
(3)
:
206
221
.
Sterio
DC
.
1984
.
The unbiased estimation of number and sizes of arbitrary particles using the disector
.
J Microsc
.
134
(Pt 2)
:
127
136
.
Stratford
K
Mason
A
Larkman
A
Major
G
Jack
J
.
1989
.
The modeling of pyramidal neurons in the visual cortex
. In:
Durbin
R
Miall
C
Mitchison
G
, editors.
The computing neuron
.
London, UK
:
, pp.
296
321
.
Tombaugh
GC
Rowe
WB
Rose
GM
.
2005
.
The slow afterhyperpolarization in hippocampal CA1 neurons covaries with spatial learning ability in aged Fisher 344 rats
.
J Neurosci
.
25
(10)
:
2609
2616
.
Trick
GL
Silverman
SE
.
1991
.
Visual sensitivity to motion: age-related changes and deficits in senile dementia of the Alzheimer type
.
Neurology
.
41
:
1437
1440
.
Vincent
SL
Peters
A
Tigges
J
.
1989
.
Effects of aging on the neurons within area 17 of rhesus monkey cerebral cortex
.
Anat Rec
.
223
(3)
:
329
341
.
Wang
XJ
Tegnér
J
Constantinidis
C
Goldman-Rakic
PS
.
2004
.
Division of labor among distinct subtypes of inhibitory neurons in a cortical microcircuit of working memory
.
.
101
(5)
:
1368
1373
.
Wang
Y
Zhou
Y
Ma
Y
Leventhal
AG
.
2005
.
Degradation of signal timing in cortical areas V1 and V2 of senescent monkeys
.
Cereb Cortex
.
15
:
403
408
.
Waters
J
Schaefer
A
Sakmann
B
.
2005
.
Backpropagating action potentials in neurones: measurement, mechanisms and potential functions
.
Prog Biophys Mol Biol
.
87
:
145
170
.
Wearne
SM
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
.
Yang
Y
Liang
Z
Li
GX
Wang
YC
Zhou
YF
.
2009
.
Aging affects response variability of V1 and MT neurons in rhesus monkeys
.
Brain Res
.
1274
:
21
27
.
Yu
S
Wang
XS
Fu
Y
Zhang
J
Ma
YY
Wang
YC
Zhou
YF
.
2005
.
Effects of age on latency and variability of visual response in monkeys
.
Chin Sci Bull
.
50
:
1163
1165
.
Zhang
J
Wang
X
Wang
Y
Fu
Y
Liang
Z
Ma
Y
Leventhal
AG
.
2008
.
Spatial and temporal sensitivity degradation of primary visual cortical cells in senescent rhesus monkeys
.
Eur J Neurosci
.
28
:
201
207
.