Diabetes can damage both the peripheral sensory organs, causing retinopathy, and the central visual system, leading to contrast sensitivity and impaired color vision in patients without retinopathy. Orientation discrimination is important for shape recognition by the visual system. Our psychophysical findings in this study show diminished orientation discrimination in patients with diabetes without retinopathy. To reveal the underlying mechanism, we established a diabetic mouse model and recorded in vivo electrophysiological data in the dorsal lateral geniculate nucleus (dLGN) and primary visual cortex (V1). Reduced orientation selectivity was observed in both individual and populations of neurons in V1 and dLGN, which increased in severity with disease duration. This diabetes-associated neuronal dysfunction appeared earlier in the V1 than dLGN. Additionally, neuronal activity and signal-to-noise ratio are reduced in V1 neurons of diabetic mice, leading to a decreased capacity for information processing by V1 neurons. Notably, the V1 in diabetic mice exhibits reduced excitatory neuronal activity and lower levels of phosphorylated mammalian target of rapamycin (mTOR). Our findings show that altered responses of both populations of and single V1 neurons may impair fine vision, thus expanding our understanding of the underlying causes of diabetes-related impairment of the central nervous system.

Diabetes has major impacts on patients worldwide, with type 2 diabetes representing the most common form of the disease, accounting for approximately 90% of all cases (1). Diabetes not only affects the peripheral system but also directly damages the central nervous system (2). Diabetic retinopathy is the leading cause of preventable blindness in working-age people (3). However, for patients with diabetes without retinopathy, the impairment of contrast sensitivity and color vision suggests damage to the visual pathway (4,5). Furthermore, a reduction in the volume of gray matter occurs in the occipital lobe of the brains of patients with type 2 diabetes (6), and low-frequency fluctuations in the occipital cortex are reduced in the resting state (7). These findings suggest abnormal neuronal function in the visual cortex; however, how visual function becomes impaired in individuals with diabetes remains unknown.

Orientation discrimination can be indispensable for shape recognition, and orientation selectivity (an important receptive field property in the primary visual cortex [V1]) provides the neural basis for orientation discrimination (8,9). We therefore investigated the cortical mechanisms of diabetes-related visual impairment through V1 neuron orientation selectivity.

To this end, we used psychophysical methods to evaluate orientation discrimination in patients with type 2 diabetes and found that the orientation discrimination of patients with diabetes was impaired. We then established a type 2 diabetic mouse model to explore the neural mechanisms contributing to this effect. Through in vivo electrophysiological methods, we observed decreased orientation selectivity in both individual and populations of neurons in the V1 and dorsal lateral geniculate nucleus (dLGN). Neuronal activity and signal-to-noise ratio (SNR) in the V1 and dLGN decreased, and damage to the dLGN was significantly delayed compared with that observed in the V1. Furthermore, phosphorylated levels of the protein kinase mammalian target of rapamycin (mTOR), which is involved in translational control and persistent synaptic plasticity and promotes neuronal survival, learning, and memory (10), were significantly reduced in the V1 of diabetic mice, suggesting its involvement in the attenuation of information processing by the V1. These results highlight the impairment of individual neurons as well as populations of neurons in the V1 as a key mechanism driving diabetes-related impairment of fine vision.

This study was approved by the Ethics Committee of the University of Science and Technology of China, and all protocols conformed to the tenets of the Declaration of Helsinki. All experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All participants provided written informed consent.

Psychophysical Tests

All participants underwent a comprehensive binocular health examination for visual diseases such as strabismus, amblyopia, cataract, and macular degeneration, performed by experienced ophthalmologists. Participants with the following conditions were excluded: cognitive impairment (with use of the recommended cutoff score of 26) (11), strabismus, anisometropia, history of eye or brain pathologies, or previous operations for these conditions. We excluded four patients with diabetes with significant fundus lesions observed with a panoramic ophthalmoscope (Daytona, P200T; Optos, Dunfermline, U.K.) and an AngioVue OCTA device (Optovue, Inc., Fremont, CA) (Supplementary Fig. 1). A total of 33 normal subjects and 34 patients with diabetes completed the test. Participants had normal or corrected-to-normal vision (not greater than 0.2 logMAR) and were naive to the purpose of the experiment (Table 1).

Table 1

Clinical details of subjects participating in the psychophysical test

Control groupGroup with diabetes
n 33 34 
Age, years   
 Mean ± SD 57.39 ± 6.66 57.38 ± 7.39 
 Median (range) 58 (41–72) 57 (42–74) 
Sex   
 Female 24 16 
 Male 18 
HbA1c   
 % 5.56 ± 0.33 8.96 ± 2.43 
 mmol/mol 37.42 ± 3.63 74.50 ± 26.46 
Duration of diabetes, years   
 Mean ± SD — 9.79 ± 5.50 
 Median (range) — 10 (1–21) 
Best corrected visual acuity, logMAR   
 OD 0.049 ± 0.061 0.052 ± 0.053 
 OS 0.026 ± 0.072 0.043 ± 0.053 
Education background   
 Primary school 15 12 
 Middle school 18 19 
 College and above 
Control groupGroup with diabetes
n 33 34 
Age, years   
 Mean ± SD 57.39 ± 6.66 57.38 ± 7.39 
 Median (range) 58 (41–72) 57 (42–74) 
Sex   
 Female 24 16 
 Male 18 
HbA1c   
 % 5.56 ± 0.33 8.96 ± 2.43 
 mmol/mol 37.42 ± 3.63 74.50 ± 26.46 
Duration of diabetes, years   
 Mean ± SD — 9.79 ± 5.50 
 Median (range) — 10 (1–21) 
Best corrected visual acuity, logMAR   
 OD 0.049 ± 0.061 0.052 ± 0.053 
 OS 0.026 ± 0.072 0.043 ± 0.053 
Education background   
 Primary school 15 12 
 Middle school 18 19 
 College and above 

Data are means ± SD, n, or median (range).

Stimuli generated with Psychtoolbox-3 (12) were displayed in the center of a 21 inch CRT screen (Sony MultiScan G520, Sony Corporation, Tokyo, Japan) (Fig. 1A). Sine grating with Gaussian blurring (Gabor) was used as the visual stimulus (Fig. 1B), and the order of appearance of each visual stimulus was controlled through a weighted up-down adaptive program (13). Participants responded to the keyboard by choosing one of two options according to the Gabor orientation relative to the vertical deflection orientations (clockwise or counterclockwise) (Fig. 1C). Target orientation started with strong values (±21°) and varied with an adaptive procedure. The clockwise response rate (p[x]) for each condition (x) was counted for each subject. For the data set (xi, p[xi]), a logistic function was used to fit the psychophysical function curve (14). Threshold, bias, lapse rate, and sensitivity index were used as parameters to measure the subjects’ psychological function (Fig. 1D and Supplementary Material).

Figure 1

Experimental design for stimuli, tasks, and psychometric curves for psychophysical tests. A: Subjects responded to different stimuli on the monitor. The subject sat at a distance of 1.5 m from the screen, with the forehead and chin remaining stationary throughout the test. In the test, the subject observed the stimuli appearing in the center of the screen with both eyes and provided responses through the keyboard. The monitor was the only light source in the test environment. B: Example stimuli consisting of sinusoidal grating with Gaussian blur (Gabor) with 90% contrast, a spatial frequency of three cycles per degree, and a visual angle of 2/3 degrees for testing orientation features. After the test starts, the vertical angle of the Gabor shifts from −21 to 21 degrees according to a weighted up-and-down adaptive procedure. Negative angles represent counterclockwise deflection of the Gabor with respect to the vertical direction. The representative stimulus shows the Gabor at an angle of −21 degrees. C: The procedure for a single task testing orientation. The subjects were asked to look at a fixed point in the center of the screen, which appears in the center of the screen after the subjects press the space bar. The subjects press a button to respond to different orientations of the Gabor stimulus. Subjects press the right arrow key when the stimulus rotates clockwise relative; they press the left arrow if the stimulus rotates counterclockwise. Subjects then press the space bar again to proceed to the next trial until the end of the test. D: Psychometric curves (Pc). Psychophysical function curves were fitted according to the clockwise rate (px = yx / nx) of the subjects for each stimulus angle (x). Bias (equal to |μ|) is equal to the horizontal distance from the midpoint of the curve to the zero point of the coordinate axis. The threshold (equal to σ) is equal to the horizontal distance from the point corresponding to the upper threshold limit of the psychometric curve to the midpoint. The lapse rate is equal to the vertical distance between the point corresponding to the maximum value of the psychometric curve and 1 (100%). The sensitivity index is the slope between the two points corresponding to 75% and 50% of the maximum value of the psychometric curve.

Figure 1

Experimental design for stimuli, tasks, and psychometric curves for psychophysical tests. A: Subjects responded to different stimuli on the monitor. The subject sat at a distance of 1.5 m from the screen, with the forehead and chin remaining stationary throughout the test. In the test, the subject observed the stimuli appearing in the center of the screen with both eyes and provided responses through the keyboard. The monitor was the only light source in the test environment. B: Example stimuli consisting of sinusoidal grating with Gaussian blur (Gabor) with 90% contrast, a spatial frequency of three cycles per degree, and a visual angle of 2/3 degrees for testing orientation features. After the test starts, the vertical angle of the Gabor shifts from −21 to 21 degrees according to a weighted up-and-down adaptive procedure. Negative angles represent counterclockwise deflection of the Gabor with respect to the vertical direction. The representative stimulus shows the Gabor at an angle of −21 degrees. C: The procedure for a single task testing orientation. The subjects were asked to look at a fixed point in the center of the screen, which appears in the center of the screen after the subjects press the space bar. The subjects press a button to respond to different orientations of the Gabor stimulus. Subjects press the right arrow key when the stimulus rotates clockwise relative; they press the left arrow if the stimulus rotates counterclockwise. Subjects then press the space bar again to proceed to the next trial until the end of the test. D: Psychometric curves (Pc). Psychophysical function curves were fitted according to the clockwise rate (px = yx / nx) of the subjects for each stimulus angle (x). Bias (equal to |μ|) is equal to the horizontal distance from the midpoint of the curve to the zero point of the coordinate axis. The threshold (equal to σ) is equal to the horizontal distance from the point corresponding to the upper threshold limit of the psychometric curve to the midpoint. The lapse rate is equal to the vertical distance between the point corresponding to the maximum value of the psychometric curve and 1 (100%). The sensitivity index is the slope between the two points corresponding to 75% and 50% of the maximum value of the psychometric curve.

Close modal

Type 2 Diabetes Mouse Model

Male C57BL/6 mice (age 7 weeks, mean ± SD weight 20 ± 2 g) were purchased from SPF Biotechnology Co. (Beijing, China) [permit no. SCXK (JING) 2019-0010]. The mice were raised according to supplier specifications and maintained under a 12 h light/dark cycle (lights on from 0700 to 1900 h) at 23–25°C. Food and water were freely available.

After a week of acclimatization, mice were randomly divided into two groups. For 8 weeks, the control group was provided standard chow and the experimental group was fed a high fat diet (HFD) (60% fat, no. TP23510; TROPHIC Animal Feed High-tech Co Ltd, Nantong, China). Type 2 diabetes was induced in the experimental mice with a combination of HFD and streptozotocin as previously described (15). Briefly, mice in the HFD group received daily injections of streptozotocin (60 mg/kg i.p.; Sigma-Aldrich, St. Louis, MO) in citrate buffer (0.1 mol/L sodium citrate and 0.1 mol/L citric acid, pH 4.5) for five consecutive days, while the standard chow group was provided the same dose of citrate buffer alone. After 1 week, fasting blood glucose (BG) concentrations were measured in each mouse with a BG meter (Accu-Chek Performa; Roche Diagnostics GmbH, Mannheim, Germany). Fasting BG levels >11 mmol/L indicated successful induction of diabetes. Subsequently, the type 2 diabetes group was continued on the HFD. Electrophysiological recordings were performed at 2 and 4 weeks after the model was established.

Oral Glucose Tolerance and Insulin Tolerance Tests

As previously described, the oral glucose tolerance test and insulin tolerance test were both performed after 8 weeks of HFD administration and at 2 weeks after the model was successfully established. Area under the curve was calculated to quantitatively evaluate insulin resistance and glucose clearance activity (Supplementary Material).

Measurement of Changes in Retinal Morphology

Retinal thickness and cell counts were quantified with ImageJ software to analyze hematoxylin-eosin (H-E)-stained paraffin section images (3 μm) (n = 6 mice per group, 3 sections per sample were calculated and averaged) as previously described (16,17). Total retinal thickness was measured as the length from the ganglion cell layer to the outer nuclear layer (ONL). Total retinal thickness, inner nuclear layer thickness, and ONL thickness measurements were made in the posterior retina at four points: two on either side of the optic nerve that were ∼200–300 μm apart (16). These measurements were then averaged to yield a measurement for that particular section. The number of cells in the ganglion cell layer was expressed in the data as number of cells per 100 μm retinal length. The measurements were taken at six points along the retina: three adjacent fields on the temporal and nasal sides beginning 200 μm from the optic nerve (17). The six measurements were averaged per eye to yield the average number of cells per animal. Measurement collection was performed with blinding as to the control and diabetic groups. All counting was performed with blinding.

In Vivo Electrophysiology

With a well-established method, linear electrodes were used to implant V1 and dLGN in anesthetized mice for in vivo electrophysiological nerve signal recording (see Supplementary Material for details). Drift grating (Fig. 2A and B) was used as a visual stimulus and displayed on a CRT monitor (Sony MultiScan G220; Sony Corporation) 30 cm from the animal. All neuronal signals were passed through a front-end amplifier (cutoff frequency 10 kHz; 1000× Blackrock Microsystems, Salt Lake City, UT) and digitized by a neural signal processing system (sampling frequency 30 kHz, 16 bits; 1000× Blackrock Microsystems); spikes were saved and detected with use of a cluster analysis for offline data analysis. Single-unit (SU) isolation was achieved via offline classification of spike waveforms with Offline-Sorter (version 3.3.5; Plexon, Dallas, TX) software (Fig. 2C and D).

Figure 2

Stimulus design and examples of spike and LFP activity of in vivo electrophysiology recordings. A: Example of a drifting grating. The stimulus was a drifting grating with 95% contrast that shifts in the direction of the yellow arrow. B: Diagram of the stimulation protocol. A blank screen with mean luminance (45.2 cd/m2) was first presented for 0.5 s. The stimulus was then displayed on the screen for 1 s, followed by a 0.5-s blank screen with mean luminance. The receptive field and optimal spatial and temporal parameters were characterized before the start of the recording. Grating under optimal parameters moved randomly in 12 different directions (0–330 degrees with an interval of 30 degrees and a blank stimulus for spontaneous activity recording). Each stimulus was presented 10 times in a pseudorandom sequence. C: Representative results of spike sorting from an electrode site recording. A principal components analysis scatterplot of spike waveforms shows three well-isolated SUs. Different colors represent different individual neurons. D: Spike waveforms of four separated SUs were isolated in comparing the spike waveforms from different sites. Different colors represent different individual neurons. E: Raw spike traces (filtered 750–4 kHz) for 10 single trials for MU activity in response to a grating stimulus (control group). Vertical dashed lines indicate onset and offset of the visual stimulus. The red solid line in each trial indicates the threshold at 3 SD of MU activity. Within each window, spike activity can be clearly seen with the initiation of stimuli. F: Raster and mean spike density plots (SEM indicated by dark-gray shading) for MU activity (control group). Light-gray shading indicates the analysis window (0.5–1.5 s). Spike activity can be clearly seen during each window of stimulation. G: LFP traces (filtered 3–300 Hz) for 10 single trials (control group), extracted from the same recording site used in E. Vertical dashed lines indicate onset and offset of visual stimuli, and gray areas indicate the analysis window (0.5–1.5 s). LFP activity can be clearly seen during the window of stimulation. H: LFP spectrograms of the response to a grating stimulus (control group). Vertical dashed lines indicate the onset and offset of visual stimulation, and horizontal solid lines indicate the γ-band (25–90 Hz). For both groups, LFP activity can be clearly seen during the stimulation period, with a prominent increase in power in the γ-band. LFP activity was detected during the window of stimulation.

Figure 2

Stimulus design and examples of spike and LFP activity of in vivo electrophysiology recordings. A: Example of a drifting grating. The stimulus was a drifting grating with 95% contrast that shifts in the direction of the yellow arrow. B: Diagram of the stimulation protocol. A blank screen with mean luminance (45.2 cd/m2) was first presented for 0.5 s. The stimulus was then displayed on the screen for 1 s, followed by a 0.5-s blank screen with mean luminance. The receptive field and optimal spatial and temporal parameters were characterized before the start of the recording. Grating under optimal parameters moved randomly in 12 different directions (0–330 degrees with an interval of 30 degrees and a blank stimulus for spontaneous activity recording). Each stimulus was presented 10 times in a pseudorandom sequence. C: Representative results of spike sorting from an electrode site recording. A principal components analysis scatterplot of spike waveforms shows three well-isolated SUs. Different colors represent different individual neurons. D: Spike waveforms of four separated SUs were isolated in comparing the spike waveforms from different sites. Different colors represent different individual neurons. E: Raw spike traces (filtered 750–4 kHz) for 10 single trials for MU activity in response to a grating stimulus (control group). Vertical dashed lines indicate onset and offset of the visual stimulus. The red solid line in each trial indicates the threshold at 3 SD of MU activity. Within each window, spike activity can be clearly seen with the initiation of stimuli. F: Raster and mean spike density plots (SEM indicated by dark-gray shading) for MU activity (control group). Light-gray shading indicates the analysis window (0.5–1.5 s). Spike activity can be clearly seen during each window of stimulation. G: LFP traces (filtered 3–300 Hz) for 10 single trials (control group), extracted from the same recording site used in E. Vertical dashed lines indicate onset and offset of visual stimuli, and gray areas indicate the analysis window (0.5–1.5 s). LFP activity can be clearly seen during the window of stimulation. H: LFP spectrograms of the response to a grating stimulus (control group). Vertical dashed lines indicate the onset and offset of visual stimulation, and horizontal solid lines indicate the γ-band (25–90 Hz). For both groups, LFP activity can be clearly seen during the stimulation period, with a prominent increase in power in the γ-band. LFP activity was detected during the window of stimulation.

Close modal

We processed the spike and local field potential (LFP) signals according to previously described procedures (18). Multiunit (MU) activity was extracted from the original data by filtering through a four-pole high pass Butterworth digital filter (cutoff frequency 300 Hz), with a threshold at 3 SD (Fig. 2E and F). For the LFP activity, the raw data were subjected to low-pass filtering and down sampled to 3 kHz. The residual 50 Hz line noise was removed. The frequency pass band for the LFP activity was 3–300 Hz. We calculated the difference in power (dB%) of the γ-band (25–90 Hz) between the appearance of a stimulus and that of a blank stimulus (Fig. 2G and H).

At one recording site (Fig. 2C), three SUs were separated, and the waveforms of the SUs were recorded at the other sites (Fig. 2D). From a single recording site, we observed significant firing (MU, 10 sweeps) in the stimulus window within the spike’s traces (Fig. 2E, Supplementary Fig. 2A, and Supplementary Video 1). This clear firing was also observed in the raster and peak density with MU activity (Fig. 2F and Supplementary Fig. 2C). Figure 2G and Supplementary Fig. 2B show the LFP traces recorded in the same 10 trials as the spikes in Fig. 2E and Supplementary Fig. 2A. Furthermore, the rapid oscillation increased obviously during the presentation of visual stimuli (Supplementary Video 2). Moreover, the LFP signal spectrogram (Fig. 2H and Supplementary Fig. 2D) showed a clear increase in the signal power of the entire γ-band (25–90 Hz). Only γ-band frequency was examined in subsequent LFP analyses.

All analytical methods were performed in the same manner for detection of SU, MU, and LFP responses, and orientation bias (OB) was calculated to measure the orientation selectivity. The neuron responses for each orientation data point were fitted with a double von Mises distribution. To ensure that the measurement was precise, we only considered the SU, MU, or LFP responses with a goodness of fit (R2) >0.8 for further analysis. The half width at half height (HWHH) and orientation selectivity index (OSI) were used to characterize the tuning curve. The SNR was used to measure the signal transmission and detection fidelity (Supplementary Material).

Western Blotting

Proteins were extracted with RIPA buffer, and 30 µg protein was separated with SDS-PAGE. The separated proteins were transferred to a polyvinylidene difluoride membrane (Merck Millipore), which was then saturated (5% nonfat dry milk or 5% BSA) and incubated with primary and secondary antibodies. Proteins were detected by enhanced chemiluminescence (Biosharp), and quantification was performed with ImageJ software (National Institutes of Health). Phosphorylated proteins were quantified for comparison with their unphosphorylated counterpart (See Supplementary Table 1 for details of the antibodies.)

Statistical Analyses

Unpaired two-tailed t tests, Mann-Whitney nonparametric tests, ANOVA analysis, or Spearman correlations were performed with SPSS, version 20. Statistical significance was set at P < 0.05.

Post hoc analyses of the samples’ statistical power and effect size were conducted with G*Power 3.1 (https://gpower.software.informer.com/3.1/) (Supplementary Tables 2 and 3). Except where noted, data are represented by the mean ± SD.

Data and Resource Availability

The data sets generated and analyzed during this study are available from the corresponding author on reasonable request.

Impaired Orientation Discrimination in Subjects With Diabetes

As the stimulus intensity decreased from strong values, the subjects’ perceptions became less reliable (Fig. 3A and B). We fitted the psychological curve for two representative subjects using a logistic function (Fig. 3C), which showed that the threshold for the subjects with diabetes was greater than that for the control subjects (Fig. 3D), meaning that the subjects with diabetes required stronger stimuli to make reliable reports (>84%). Compared with control participants, the perceived reference point (midpoint) of the patients with diabetes showed greater bias from the true midpoint (0°), although the differences were not statistically significant (Fig. 3E). The curve maximum for the subjects with diabetes was smaller than that of the control group, indicating a greater judgment lapse rate for participants with diabetes than for control subjects (Fig. 3F). Additionally, the subjects with diabetes had lower sensitivity index scores than control subjects (Fig. 3G), suggesting a weaker perception of changes in orientation compared with that of healthy individuals.

Figure 3

Representative example of the orientation discrimination test and statistical results (A and B). Example of psychometric function measurement in the orientation task (left, control subject; right, subject with diabetes). According to the program, the two staircases were assigned up or down steps of 5/2 and 2/5 degrees, with convergence points of 71.43% and 28.57%, respectively. In both staircases, the next stimulus increases the number of steps after a correct response or decreases the steps after incorrect answers. The distribution of dots for the group without diabetes deviated more from the vertical orientation than that for the control group. Negative degrees indicate a counterclockwise shift from vertical orientation (0°). The black dashed line represents vertical orientation. C: Pooled data from each target orientation level (dots) reached by a subject, and the corresponding Bayesian estimate of the psychometric function (curve). Example of control curve parameters (LCX, dashed curve): threshold = 1.28, bias = 0.86, lapse rate = 0.3%, sensitivity index = 0.31. Example of diabetes curve parameters (LSM, solid curve): threshold = 3.83, bias = 1.59, lapse rate = 1.1%, sensitivity index = 0.10. The square mark indicates the midpoint of the curve. DG: Statistical analysis showing impaired orientation discrimination in subjects without diabetes. Combination histogram and scatterplot depicting comparison between group with diabetes and control groups; bias between the two groups was not significantly different (E), and the group with diabetes had a higher threshold (D), higher lapse rate (F), and lower sensitivity index (G). H: There was a significant positive correlation between the log lapse rate and the glycosylated hemoglobin levels. I: In patients with poorly controlled glycosylated hemoglobin levels (HbA1c ≥8%), log lapse rate was significantly positively correlated with disease duration. Bar graphs show mean ± SD. ns, no significant difference. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by Mann-Whitney U test.

Figure 3

Representative example of the orientation discrimination test and statistical results (A and B). Example of psychometric function measurement in the orientation task (left, control subject; right, subject with diabetes). According to the program, the two staircases were assigned up or down steps of 5/2 and 2/5 degrees, with convergence points of 71.43% and 28.57%, respectively. In both staircases, the next stimulus increases the number of steps after a correct response or decreases the steps after incorrect answers. The distribution of dots for the group without diabetes deviated more from the vertical orientation than that for the control group. Negative degrees indicate a counterclockwise shift from vertical orientation (0°). The black dashed line represents vertical orientation. C: Pooled data from each target orientation level (dots) reached by a subject, and the corresponding Bayesian estimate of the psychometric function (curve). Example of control curve parameters (LCX, dashed curve): threshold = 1.28, bias = 0.86, lapse rate = 0.3%, sensitivity index = 0.31. Example of diabetes curve parameters (LSM, solid curve): threshold = 3.83, bias = 1.59, lapse rate = 1.1%, sensitivity index = 0.10. The square mark indicates the midpoint of the curve. DG: Statistical analysis showing impaired orientation discrimination in subjects without diabetes. Combination histogram and scatterplot depicting comparison between group with diabetes and control groups; bias between the two groups was not significantly different (E), and the group with diabetes had a higher threshold (D), higher lapse rate (F), and lower sensitivity index (G). H: There was a significant positive correlation between the log lapse rate and the glycosylated hemoglobin levels. I: In patients with poorly controlled glycosylated hemoglobin levels (HbA1c ≥8%), log lapse rate was significantly positively correlated with disease duration. Bar graphs show mean ± SD. ns, no significant difference. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by Mann-Whitney U test.

Close modal

Analysis of Spearman correlation coefficients between diabetes duration and glycosylated hemoglobin and four parameters (log threshold, log bias, log lapse rate, and log sensitivity index) showed no significant correlations between duration of diabetes and any of the four parameters (Supplementary Fig. 3AD). However, lapse rate was significantly positively correlated with glycosylated hemoglobin level (Fig. 3H and Supplementary Fig. 3EH). In addition, we found a significant positive correlation between the log lapse rate and disease duration for patients with poorly controlled glycosylated hemoglobin levels (HbA1c ≥8%) (Fig. 3I and Supplementary Fig. 3IQ).

Decreased Orientation Selectivity in the V1 and dLGN of Diabetic Mice

We then established a diabetic mouse model to investigate whether diabetes damages the visual center. The diabetic mice showed sustained hyperglycemia, glucose intolerance, and insulin resistance consistent with clinical manifestations of diabetes (Supplementary Material and Supplementary Fig. 4).

For assessment of effects of diabetes on the retina, we used H-E staining to evaluate retinal morphology in diabetic mice at 2 and 4 weeks of observation (Supplementary Fig. 5). Ganglion cell counts, total retinal thickness, ONL, and inner nuclear layer of diabetic mice did not change significantly compared with the control mice at 2 or 4 weeks after diabetes induction.

We performed extracellular in vivo electrophysiological recordings of the V1 and dLGN neurons in 32 diabetic mice and 34 control mice (Supplementary Table 4). Spike activity and LFPs were also recorded.

We quantitatively evaluated the orientation selectivity of the SU, MU, and LFP signals in the V1 of the mice and generated orientation tuning curves of the SUs in the V1 of control and diabetic mice (Figs. 4A and 4B). The MU and LFP responses were recorded at the same site (Fig. 4D, E, G, and H). The MU and LFP signals clearly showed orientation tuning to grating stimuli and had optimal tuning orientations similar to those of SUs recorded at the same site. The SU orientation selectivity of the diabetic mice (at both 2 and 4 weeks post–onset of diabetes) was weaker than that of the control mice (Fig. 4C), as were the MU and LFP (Fig. 4F and I).

Figure 4

Representative examples of orientation tuning curves and statistical comparisons. A and B: Orientation tuning curves for isolated SU recorded from one electrode site (control, A, and diabetes, B). OB was used to measure the orientation selectivity (optimal tuning orientation [Opt.Ori]). C: The SU orientation selectivity in diabetic and control mice. Since there was no significant difference in OB values between the 2-week and 4-week control groups, the data from those groups were combined (data not shown), as were the (F) and (I). D and E: MU tuning curve extracted from the same recording site as for A and B. F: The orientation selectivity of MU in diabetic and control mice. G and H: A γ LFP tuning curve extracted from the same recording site as for A and B. I: The orientation selectivity of LFP in diabetic and control mice. J and K: Tuning curves were constructed for an SU by fitting the responses for each task to double von Mises function. Solid dots represent neuronal responses to different orientations, error bars represent SD, solid lines represent fitted curves, shaded areas indicate ± 1 SD, and J and K show the SU recorded in the control and diabetes groups, respectively, estimated by bootstrapping (nboot = 1,000). L and M: The cumulative frequency distribution curve and the bar graph show the statistical analysis of bandwidth (L) and OSIs (M) of the SU tuning curve for the control and diabetic groups. HWHH was used to characterize bandwidth. N and O: Tuning curves were constructed for the MU by fitting the responses of control (N) and diabetic (O) groups for each task to a double von Mises function. Solid dots represent neuronal responses to different orientations, error bars represent SD, solid lines represent fitted curves, and shaded areas indicate ± 1 SD, estimated by bootstrapping (nboot = 1,000). P and Q: The cumulative frequency distribution curve and the bar graph show the statistical results of the HWHH (P) and the OSI (Q) of the MU tuning curve for the control and diabetic groups. R and S: We constructed Tuning curves for the LFP by fitting the responses by control (R) and diabetes (S) for each task to a double von Mises distribution. Solid dots represent LFP responses to different orientations, solid lines represent fitted curves, and shaded areas indicate ±1 SD, estimated by bootstrapping (nboot = 1,000). T and U: Curve of the cumulative frequency distribution and bar graph showing statistical analysis of the HWHH (T) and OSI (U) of the LFP tuning curve for the control and diabetes groups. Bar graphs show mean ± SD. ns, no significant difference. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by one-way ANOVA with Tukey multiple comparisons test (C, F, and I) or Mann-Whitney U test (L, M, P, Q, T, and U). spk/sec, spikes per second.

Figure 4

Representative examples of orientation tuning curves and statistical comparisons. A and B: Orientation tuning curves for isolated SU recorded from one electrode site (control, A, and diabetes, B). OB was used to measure the orientation selectivity (optimal tuning orientation [Opt.Ori]). C: The SU orientation selectivity in diabetic and control mice. Since there was no significant difference in OB values between the 2-week and 4-week control groups, the data from those groups were combined (data not shown), as were the (F) and (I). D and E: MU tuning curve extracted from the same recording site as for A and B. F: The orientation selectivity of MU in diabetic and control mice. G and H: A γ LFP tuning curve extracted from the same recording site as for A and B. I: The orientation selectivity of LFP in diabetic and control mice. J and K: Tuning curves were constructed for an SU by fitting the responses for each task to double von Mises function. Solid dots represent neuronal responses to different orientations, error bars represent SD, solid lines represent fitted curves, shaded areas indicate ± 1 SD, and J and K show the SU recorded in the control and diabetes groups, respectively, estimated by bootstrapping (nboot = 1,000). L and M: The cumulative frequency distribution curve and the bar graph show the statistical analysis of bandwidth (L) and OSIs (M) of the SU tuning curve for the control and diabetic groups. HWHH was used to characterize bandwidth. N and O: Tuning curves were constructed for the MU by fitting the responses of control (N) and diabetic (O) groups for each task to a double von Mises function. Solid dots represent neuronal responses to different orientations, error bars represent SD, solid lines represent fitted curves, and shaded areas indicate ± 1 SD, estimated by bootstrapping (nboot = 1,000). P and Q: The cumulative frequency distribution curve and the bar graph show the statistical results of the HWHH (P) and the OSI (Q) of the MU tuning curve for the control and diabetic groups. R and S: We constructed Tuning curves for the LFP by fitting the responses by control (R) and diabetes (S) for each task to a double von Mises distribution. Solid dots represent LFP responses to different orientations, solid lines represent fitted curves, and shaded areas indicate ±1 SD, estimated by bootstrapping (nboot = 1,000). T and U: Curve of the cumulative frequency distribution and bar graph showing statistical analysis of the HWHH (T) and OSI (U) of the LFP tuning curve for the control and diabetes groups. Bar graphs show mean ± SD. ns, no significant difference. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by one-way ANOVA with Tukey multiple comparisons test (C, F, and I) or Mann-Whitney U test (L, M, P, Q, T, and U). spk/sec, spikes per second.

Close modal

HWHH of the SU tuning curve in the V1 of diabetic mice was wider than that of the control mice curve (Fig. 4J, K, N, O, R, and S), and the OSI was lower than that of control mice (Fig. 4L and M). Additionally, the tuning bandwidth and OSI of the local population signals (MU, LFP) showed similar results; diabetic mice had a higher HWHH and lower OSI compared with control mice (Fig. 4P, Q, T, and U).

In the visual pathway, the visual cortex receives direct projections from dLGN neurons, and changes in LGN neuron function affect V1 function (19,20) dLGN orientation selectivity did not change significantly in the 2-week observation period for the diabetes group but significantly decreased over 4 weeks of observation in these mice (Supplementary Fig. 6).

Diabetes-Induced Changes in the Response Properties of V1 and dLGN Neurons

There was a significant decrease in various responses of V1 neurons, including the average response in all orientations (Fig. 5A) and peak response (Fig. 5B). However, spontaneous activity was not significantly different (Fig. 5C). There was an evident decrease in the SNR values of the V1 neurons in diabetic mice (Fig. 5D). Although V1 neuronal activity was reduced in diabetic mice at 2 weeks after diabetes onset, we observed no relevant changes in dLGN neurons during the same period, whereas the average and peak responses were significantly decreased in the dLGN of diabetic mice at 4 weeks post–diabetes induction (Supplementary Fig. 7).

Figure 5

Investigation of the mechanism of diabetes affecting V1 neuron function. Average evoked firing rate (A), peak response (B), spontaneous activity (C), and SNR for the control and diabetes groups (D). E: Representative example of V1 neuron spike waveform. Duration represents the time interval between the spike trough and the peak. The end slope indicates the slope of the waveform at 0.5 ms after the trough. F: Representative FS neurons and RS neurons, separated by their spike wave characteristics. FS neurons had a shorter trough-peak duration (<0.45 ms), larger peak-to-trough ratio (>0.8), and negative end slope (<0). G and H: Diabetes was associated with decreased peak evoked firing rates of RS neurons (G) without significantly altering the peak response of FS neurons (H). I and J: Diabetes was associated with decreased orientation selectivity of RS neurons (I) and FS neurons (J). Phosphorylated (p-)mTOR, total (t-)mTOR, and β-actin proteins in V1 tissues detected by Western blot (K) and quantified by densitometric analysis (L). Bar graphs show mean ± SD. Average, average evoked response; ns, no significant difference; peak, peak response; spontaneous, spontaneous activity. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by one-way ANOVA with Tukey multiple comparisons test (AD and GJ) or two-way ANOVA with Tukey multiple comparisons test (L). ns, no significant difference.

Figure 5

Investigation of the mechanism of diabetes affecting V1 neuron function. Average evoked firing rate (A), peak response (B), spontaneous activity (C), and SNR for the control and diabetes groups (D). E: Representative example of V1 neuron spike waveform. Duration represents the time interval between the spike trough and the peak. The end slope indicates the slope of the waveform at 0.5 ms after the trough. F: Representative FS neurons and RS neurons, separated by their spike wave characteristics. FS neurons had a shorter trough-peak duration (<0.45 ms), larger peak-to-trough ratio (>0.8), and negative end slope (<0). G and H: Diabetes was associated with decreased peak evoked firing rates of RS neurons (G) without significantly altering the peak response of FS neurons (H). I and J: Diabetes was associated with decreased orientation selectivity of RS neurons (I) and FS neurons (J). Phosphorylated (p-)mTOR, total (t-)mTOR, and β-actin proteins in V1 tissues detected by Western blot (K) and quantified by densitometric analysis (L). Bar graphs show mean ± SD. Average, average evoked response; ns, no significant difference; peak, peak response; spontaneous, spontaneous activity. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by one-way ANOVA with Tukey multiple comparisons test (AD and GJ) or two-way ANOVA with Tukey multiple comparisons test (L). ns, no significant difference.

Close modal

Excitatory/Inhibitory Balance of V1 Is Impaired in Diabetic Mice

By dividing V1 neurons into fast-spiking (FS) (primarily inhibitory) neurons and regular-spiking (RS) (primarily excitatory) neurons, we could then investigate the influence of the excitatory/inhibitory balance on neuronal orientation selectivity in the V1 of diabetic mice (21). This division was based on a shorter trough-peak duration (<0.45 ms), larger peak-to-trough ratio (>0.8), and negative end slope (<0) (Fig. 5E and F). In comparisons with the control group, the RS neuron activity was lower in the diabetic group, although FS neuron activity did not change significantly (at both 2 and 4 weeks post–onset of diabetes) (Fig. 5G and H).

Moreover, at 2 weeks post–onset of diabetes, RS neurons but not FS neurons showed a significant difference in orientation selectivity between the diabetic and control groups, while at 4 weeks post–onset of diabetes, both FS and RS neurons showed reduced orientation selectivity in the diabetic group compared with the control group (Fig. 5I and J).

Measurement of V1 Tissue in Diabetic Mice

To investigate the mechanisms underlying the reduced excitability and impaired capacity for information processing in V1 neurons of diabetic mice, we examined differences in the V1 tissue between diabetic and nondiabetic mice. Glutamate and γ-aminobutyric acid (GABA) are the major excitatory and inhibitory neurotransmitters, respectively, in the mammalian cerebral cortex (22). Using high-performance liquid chromatography coupled with tandem mass spectrometry, we quantified GABA and glutamate levels in V1 tissue (see Supplementary Material for detailed methods) and found no difference in the levels of GABA and glutamate between the diabetes and control groups (Supplementary Fig. 8F and G). Moreover, the expression of GluA2, GluA3, GluN1, and GluN2A excitatory receptors were not significantly altered (Supplementary Fig. 8AE). However, Western blot analysis of phosphorylated mTOR levels showed a significant decrease in the V1 of diabetic mice compared with control animals at 4 weeks after diabetes onset (Fig. 5K and L).

In this study, we explored the cortical mechanisms of fine visual impairment in patients with type 2 diabetes. Our major findings are as follows: 1) orientation discrimination is impaired in patients with diabetes without retinopathy; 2) orientation selectivity is reduced in both individual neurons and populations of neurons in the V1 and dLGN of diabetic mice; 3) activity and SNR are both reduced in the dLGN and V1 neurons of diabetic mice; 4) V1 impairment is observed before that of the dLGN of diabetic mice, although damage appears to be at least partially inherited from dLGN; and 5) disruption of the balance between excitation/inhibition signaling and aberrant mTOR signaling may contribute to V1 neuronal dysfunction. In summary, impairment in the orientation selectivity and information processing capacity of V1 neurons is the neural basis of fine visual impairment in patients with diabetes.

The LFP, MU, and SU signals of V1 and dLGN neurons were well tuned to the grating, and SU, MU, and LFP at the same site had similar optimal tuning orientations, which is consistent with previous reports (18).

Additionally, electrophysiological studies revealed decreased orientation selectivity of both single neurons and neuron populations in diabetic mice. We calculated a median OB (V1) slightly lower than that previously reported (23), while the median OB (LGN) was similar to previous reports (24). One possible explanation for these observations is that the orientation selectivity of mouse V1 neurons (but not dLGN neurons) is contrast dependent (25), and we designed the visual stimulus with 95% contrast.

Moreover, signals provided by individual neurons form the physiological basis for information transmission throughout the nervous system (26,27). Furthermore, MU activity and LFP are thought to be the sum of responses from many neurons close to the recording site (28,29), and these responses underlie cognition and behavior (30,31), therefore suggesting that reduced orientation selectivity by individual neurons or neuron populations in the V1 could underlie the pathophysiology of impaired orientation discrimination in patients with diabetes.

In the visual nervous system, dLGN neurons (“relay neurons”) transmit information from the retina to the visual cortex to play a crucial role in the formation of V1 neuron orientation selectivity (19,20). We also found reduced orientation selectivity in the dLGN of diabetic mice at 4 weeks post–diabetes induction, suggesting that damage to orientation selectivity in V1 neurons is partially inherited from the dLGN, and interestingly, this may be initiated before damage in the dLGN. Additionally, since anesthetized mice were used for electrophysiological experiments, the top-down effects of higher cortices on orientation selectivity were excluded (32).

V1 neurons perform orientation tuning in response to a combination of excitatory and inhibitory inputs (33). It is generally accepted that the V1 receives excitatory inputs from the dLGN and other cortical cells to develop orientation selectivity (19,34), while the capacity for orientation tuning is sharpened with the help of intracortical inhibitory mechanisms (35). We found that the V1 neurons in diabetic mice showed lower peak and average evoked responses. We hypothesized that these observations were due to reduced excitatory input and increased inhibitory input received by V1 neurons. However, increasing the inhibition level reduces spontaneous activity and peak response (36), and we detected no significant changes in spontaneous activity, implying that a decrease in excitatory inputs is responsible for the decrease in peak and average evoked responses. Moreover, the reduction in peak response (signal) without significant changes in the spontaneous activity (noise) results in reduced SNR. These results thus suggest that V1 neurons in diabetic mice have attenuated capacity for information processing and reduced fidelity in their information transmission, potentially leading to impaired visual performance (37).

Balance in excitatory/inhibitory signal transmission is essential for nervous system function (22), and our results imply that reduced excitatory activity leads to reduced evoked responses and weakened orientation selectivity by V1 neurons. In addition, we found that the peak and average evoked responses of dLGN neurons are also decreased in diabetic mice, indicating that the V1 may be affected by the dLGN through decreased excitatory input. The V1 inhibitory system in patients with diabetes may be maintained in a relatively stable state through compensatory effects, such as altering the expression of GABA transporters (38) or altering the number of GABA receptors (39). Although our results showed no significant changes in the expression of excitatory (glutamate) or inhibitory (GABA) transmitters and glutamate receptors (GluA2, GluA3, GluN1, and GluN2A) in the V1 of the group with diabetes compared with control subjects, the levels of several other glutamate receptor subunits are unknown. Furthermore, the total protein levels of receptors are not necessarily representative of the levels of receptors expressed on membranes (40). Thus, in our future work we will explore how diabetes affects the cortical excitatory system.

Diabetes can impair neuronal function through multiple pathways. The protein kinase mTOR not only mediates neural processes such as learning and memory formation but also is involved in the regulation of neuronal excitability (10,41). Given the broad range of changes induced by altered mTOR signaling, however, the mechanisms underlying cortical damage in diabetes remain uncertain. Insulin receptor signaling also plays multiple roles in the brain, participating in the regulation of synaptic plasticity, neuronal survival, learning and memory, and neurological disorders (42). Notably, mTOR exhibits cross talk with the insulin receptor signaling pathway, thereby affecting neuronal survival, differentiation, and development by interfering with insulin signaling (43). Furthermore, the persistent hyperglycemic state of diabetes can, directly and indirectly, damage neurons through oxidative stress, inflammation, and apoptosis, leading to cognitive dysfunction (44,45). These findings suggest that, in diabetes, neurons have a reduced capacity for information processing, possibly attributable to a combination of aberrant downregulation of mTOR levels, impaired insulin signaling, and the effects of hyperglycemia. Future studies are necessary to explore whether insulin or antidiabetes drug treatment can rescue damage to V1 neurons.

The retinopathy phenotype observed in diabetic mice has been reported to include an increased number of astrocytes and glial cells at 4–5 weeks after hyperglycemic episodes (46), while morphological abnormalities, loss of retinal ganglion cells, and reduction in retinal thickness require 10 weeks or more after diabetes onset (16,47). H-E staining showed no significant morphological changes in the retinas of mice at either 2 weeks or 4 weeks after diabetes onset. Moreover, the orientation selectivity and response properties of dLGN neurons were not significantly altered in diabetic mice after 2 weeks of observation, suggesting that the pathway for visual information transmission from the retina to the dLGN was fully functional. However, at this point, the V1 has already shown functional impairment before peripheral damage. At 4 weeks of observation, the dLGN neurons show a decrease in orientation selectivity and reduced responses, potentially exacerbating damage to V1 neuron function. Thus, substantially longer observation periods are necessary to account for long-term changes in visual function associated with diabetes.

Aging and diabetes independently affect cognition (44,48). Moreover, the risk of cognitive dysfunction is 1.5 times higher in individuals with diabetes than in normal individuals (49), and the risk of cognitive impairment and developing dementia is increased by the combined effects of diabetes and aging (50). Previous psychophysical studies in humans and electrophysiology in macaques showed that aging affects orientation discrimination (51). Our psychophysical control group was recruited from elderly subjects who were age matched to the group with diabetes, and the psychophysical test results showed that orientation discrimination was impaired in subjects with diabetes. These results, therefore, suggest that aging and diabetes may synergistically exacerbate impairment of orientation discrimination. Our electrophysiological experiments also showed reduced V1 neuron orientation selectivity in diabetic mice. However, the mice were not old enough to reflect the dual effects of aging and diabetes on the V1, and thus continued study is necessary for better understanding of how diabetes leads to a decline in visual processing during aging.

In conclusion, these results show impairment of orientation discrimination in fine visual perception among patients with type 2 diabetes. Additionally, we provide electrophysiological evidence of reduced orientation selectivity in both single neurons and neuron populations in a mouse model of diabetes. We hypothesize that an imbalance in cortical excitatory versus inhibitory signaling in patients with diabetes is a possible mechanism driving the impairment of orientation discrimination. Moreover, aberrant downregulation of mTOR signaling may cause a decline in the capacity for information processing by V1 neurons. Our findings provide new evidence of damage to function in the fine visual discrimination of patients with diabetes and provide mechanistic insights to guide further study of the negative impacts of diabetes in the V1 and other brain regions.

This article contains supplementary material online at https://doi.org/10.2337/figshare.19807627.

H.C. and M.W. contributed equally to this work.

Acknowledgments. The authors thank Yupeng Yang and Yi Cao for their assistance during the writing of the manuscript, Ziheng Zhou (Chinese Academy of Sciences (CAS) Key Laboratory of Nutrition, Metabolism and Food Safety, Innovation Center for Intervention of Chronic Disease and Promotion of Health, Shanghai Institute of Nutrition and Health, University of Chinese Academy of Sciences, Chinese Academy of Sciences) and Yueyue Zhang (School of Life Sciences, University of Science and Technology of China) for the technical guidance for protein immunoblotting, Qilun Zhang (School of Life Sciences, University of Science and Technology of China) for the construction of a mouse model of diabetes, and Yuchong Han (School of Life Sciences, University of Science and Technology of China) and Ying Chen (School of Art, Anhui Polytechnic University) for assistance in producing the sample videos and pictures. The authors thank all participants for their involvement in the study.

Funding. This work was supported by the National Natural Science Foundation of China (grant NSFC 32070990 to Y.Z.) and Anhui Province’s Key Research and Development Plan (grant 9021033204 to L.F.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. H.C. and G.X. designed the research, and H.C. wrote the MATLAB script. L.X. and J.D. performed all the psychophysical tests. H.C., M.W., and Z.W. constructed the diabetic mouse model, and H.C. and M.W. performed the mouse electrophysiological experiments and processed the data. H.C. and Y.Z. wrote the manuscript, with contributions and comments from all of the authors. L.F. and Y.Z. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

1.
Saeedi
P
,
Petersohn
I
,
Salpea
P
, et al.;
IDF Diabetes Atlas Committee
.
Global and regional diabetes prevalence estimates for 2019 and projections for 2030 and 2045: results from the International Diabetes Federation Diabetes Atlas, 9th edition
.
Diabetes Res Clin Pract
2019
;
157
:
107843
2.
Biessels
GJ
,
Reijmer
YD
.
Brain changes underlying cognitive dysfunction in diabetes: what can we learn from MRI?
Diabetes
2014
;
63
:
2244
2252
3.
Wong
TY
,
Cheung
CM
,
Larsen
M
,
Sharma
S
,
Simó
R
.
Diabetic retinopathy
.
Nat Rev Dis Primers
2016
;
2
:
16012
4.
Safi
S
,
Rahimi
A
,
Raeesi
A
, et al
.
Contrast sensitivity to spatial gratings in moderate and dim light conditions in patients with diabetes in the absence of diabetic retinopathy
.
BMJ Open Diabetes Res Care
2017
;
5
:
e000408
5.
Feitosa-Santana
C
,
Paramei
GV
,
Nishi
M
,
Gualtieri
M
,
Costa
MF
,
Ventura
DF
.
Color vision impairment in type 2 diabetes assessed by the D-15d test and the Cambridge Colour Test
.
Ophthalmic Physiol Opt
2010
;
30
:
717
723
6.
Zhang
Y
,
Zhang
X
,
Zhang
J
, et al
.
Gray matter volume abnormalities in type 2 diabetes mellitus with and without mild cognitive impairment
.
Neurosci Lett
2014
;
562
:
1
6
7.
Cui
Y
,
Jiao
Y
,
Chen
Y-C
, et al
.
Altered spontaneous brain activity in type 2 diabetes: a resting-state functional MRI study
.
Diabetes
2014
;
63
:
749
760
8.
Zhan
X
,
Shou
T
.
Anatomical evidence of subcortical contributions to the orientation selectivity and columns of the cat’s primary visual cortex
.
Neurosci Lett
2002
;
324
:
247
251
9.
Hubel
DH
,
Wiesel
TN
.
Receptive fields of single neurones in the cat’s striate cortex
.
J Physiol
1959
;
148
:
574
591
10.
Hoeffer
CA
,
Klann
E
.
mTOR signaling: at the crossroads of plasticity, memory and disease
.
Trends Neurosci
2010
;
33
:
67
75
11.
Hu
JB
,
Zhou
WH
,
Hu
SH
, et al
.
Cross-cultural difference and validation of the Chinese version of Montreal Cognitive Assessment in older adults residing in Eastern China: preliminary findings
.
Arch Gerontol Geriatr
2013
;
56
:
38
43
12.
Brainard
DH
.
The psychophysics toolbox
.
Spat Vis
1997
;
10
:
433
436
13.
Kaernbach
C
.
Simple adaptive testing with the weighted up-down method
.
Percept Psychophys
1991
;
49
:
227
229
14.
Treutwein
B
,
Strasburger
H
.
Fitting the psychometric function
.
Percept Psychophys
1999
;
61
:
87
106
15.
Zhu
Y
,
Zhu
C
,
Yang
H
,
Deng
J
,
Fan
D
.
Protective effect of ginsenoside Rg5 against kidney injury via inhibition of NLRP3 inflammasome activation and the MAPK signaling pathway in high-fat diet/streptozotocin-induced diabetic mice
.
Pharmacol Res
2020
;
155
:
104746
16.
Martin
PM
,
Roon
P
,
Van Ells
TK
,
Ganapathy
V
,
Smith
SB
.
Death of retinal neurons in streptozotocin-induced diabetic mice
.
Invest Ophthalmol Vis Sci
2004
;
45
:
3330
3336
17.
Navneet
S
,
Zhao
J
,
Wang
J
, et al
.
Hyperhomocysteinemia-induced death of retinal ganglion cells: the role of Müller glial cells and NRF2
.
Redox Biol
2019
;
24
:
101199
18.
Li
G
,
Baker
CL
Jr
.
Functional organization of envelope-responsive neurons in early visual cortex: organization of carrier tuning properties
.
J Neurosci
2012
;
32
:
7538
7549
19.
Hubel
DH
,
Wiesel
TN
.
Receptive fields, binocular interaction and functional architecture in the cat’s visual cortex
.
J Physiol
1962
;
160
:
106
154
20.
Kondo
S
,
Ohki
K
.
Laminar differences in the orientation selectivity of geniculate afferents in mouse primary visual cortex
.
Nat Neurosci
2016
;
19
:
316
319
21.
Azouz
R
,
Gray
CM
,
Nowak
LG
,
McCormick
DA
.
Physiological properties of inhibitory interneurons in cat striate cortex
.
Cereb Cortex
1997
;
7
:
534
545
22.
Petroff
OA
.
GABA and glutamate in the human brain
.
Neuroscientist
2002
;
8
:
562
573
23.
Zhao
X
,
Chen
H
,
Liu
X
,
Cang
J
.
Orientation-selective responses in the mouse lateral geniculate nucleus
.
J Neurosci
2013
;
33
:
12751
12763
24.
Piscopo
DM
,
El-Danaf
RN
,
Huberman
AD
,
Niell
CM
.
Diverse visual features encoded in mouse lateral geniculate nucleus
.
J Neurosci
2013
;
33
:
4642
4656
25.
Dai
WP
,
Zhou
D
,
McLaughlin
DW
,
Cai
D
.
Mechanisms underlying contrast-dependent orientation selectivity in mouse V1
.
Proc Natl Acad Sci USA
2018
;
115
:
11619
11624
26.
Bradley
A
,
Skottun
BC
,
Ohzawa
I
,
Sclar
G
,
Freeman
RD
.
Visual orientation and spatial frequency discrimination: a comparison of single neurons and behavior
.
J Neurophysiol
1987
;
57
:
755
772
27.
Hua
T
,
Bao
P
,
Huang
CB
, et al
.
Perceptual learning improves contrast sensitivity of V1 neurons in cats
.
Curr Biol
2010
;
20
:
887
894
28.
Xing
D
,
Yeh
C-I
,
Shapley
RM
.
Spatial spread of the local field potential and its laminar variation in visual cortex
.
J Neurosci
2009
;
29
:
11540
11549
29.
Mitzdorf
U
.
Current source-density method and application in cat cerebral cortex: investigation of evoked potentials and EEG phenomena
.
Physiol Rev
1985
;
65
:
37
100
30.
Buzsáki
G
.
Large-scale recording of neuronal ensembles
.
Nat Neurosci
2004
;
7
:
446
451
31.
Fries
P
.
Neuronal gamma-band synchronization as a fundamental process in cortical computation
.
Annu Rev Neurosci
2009
;
32
:
209
224
32.
Maunsell
JH
,
Cook
EP
.
The role of attention in visual processing
.
Philos Trans R Soc Lond B Biol Sci
2002
;
357
:
1063
1072
33.
Sillito
AM
.
Inhibitory mechanisms influencing complex cell orientation selectivity and their modification at high resting discharge levels
.
J Physiol
1979
;
289
:
33
53
34.
Hoffman
KP
,
Stone
J
.
Conduction velocity of afferents to cat visual cortex: a correlation with cortical receptive field properties
.
Brain Res
1971
;
32
:
460
466
35.
Shapley
R
,
Hawken
M
,
Ringach
DL
.
Dynamics of orientation selectivity in the primary visual cortex and the importance of cortical inhibition
.
Neuron
2003
;
38
:
689
699
36.
Leventhal
AG
,
Wang
Y
,
Pu
M
,
Zhou
Y
,
Ma
Y
.
GABA and its agonists improved visual cortical function in senescent monkeys
.
Science
2003
;
300
:
812
815
37.
Ren
Z
,
Zhou
J
,
Yao
Z
, et al
.
Neuronal basis of perceptual learning in striate cortex
.
Sci Rep
2016
;
6
:
24769
38.
Baptista
FI
,
Gaspar
JM
,
Cristóvão
A
,
Santos
PF
,
Köfalvi
A
,
Ambrósio
AF
.
Diabetes induces early transient changes in the content of vesicular transporters and no major effects in neurotransmitter release in hippocampus and retina
.
Brain Res
2011
;
1383
:
257
269
39.
Wan
Q
,
Xiong
ZG
,
Man
HY
, et al
.
Recruitment of functional GABA(A) receptors to postsynaptic domains by insulin
.
Nature
1997
;
388
:
686
690
40.
Fan
J
,
Li
D
,
Chen
H-S
, et al
.
Metformin produces anxiolytic-like effects in rats by facilitating GABAA receptor trafficking to membrane
.
Br J Pharmacol
2019
;
176
:
297
316
41.
Lasarge
CL
,
Danzer
SC
.
Mechanisms regulating neuronal excitability and seizure development following mTOR pathway hyperactivation
.
Front Mol Neurosci
2014
;
7
:
18
42.
Chiu
S-L
,
Cline
HT
.
Insulin receptor signaling in the development of neuronal structure and function
.
Neural Dev
2010
;
5
:
7
43.
Garelick
MG
,
Kennedy
BK
.
TOR on the brain
.
Exp Gerontol
2011
;
46
:
155
163
44.
Rom
S
,
Zuluaga-Ramirez
V
,
Gajghate
S
, et al
.
Hyperglycemia-driven neuroinflammation compromises BBB leading to memory loss in both diabetes mellitus (DM) type 1 and type 2 mouse models
.
Mol Neurobiol
2019
;
56
:
1883
1896
45.
Klein
JP
,
Waxman
SG
.
The brain in diabetes: molecular changes in neurons and their implications for end-organ damage
.
Lancet Neurol
2003
;
2
:
548
554
46.
Kumar
S
,
Zhuo
L
.
Longitudinal in vivo imaging of retinal gliosis in a diabetic mouse model
.
Exp Eye Res
2010
;
91
:
530
536
47.
Mishra
M
,
Duraisamy
AJ
,
Kowluru
RA
.
Sirt1: a guardian of the development of diabetic retinopathy
.
Diabetes
2018
;
67
:
745
754
48.
Fischer
ME
,
Cruickshanks
KJ
,
Schubert
CR
, et al
.
Age-related sensory impairments and risk of cognitive impairment
.
J Am Geriatr Soc
2016
;
64
:
1981
1987
49.
Cukierman
T
,
Gerstein
HC
,
Williamson
JD
.
Cognitive decline and dementia in diabetes--systematic overview of prospective observational studies
.
Diabetologia
2005
;
48
:
2460
2469
50.
Lee
SM
,
Koh
D
,
Fun
SN
,
Sum
CF
.
Diabetes management and hyperglycemia in safety sensitive jobs
.
Saf Health Work
2011
;
2
:
380
384
51.
Xia
L
,
Chen
H
,
Dong
J
,
Luo
S
,
Feng
L
.
Decline of orientation and direction sensitivity in the aging population
.
Front Neurosci
2021
;
15
:
643414
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/journals/pages/license.