Chronic functional and structural imaging of the ferret primary visual cortex in vivo. (A) OD map of primary visual cortex in a ferret measured by intrinsic signal optical imaging (at 630 nm). Pixels from black to white represent the level of contralateral to ipsilateral eye drive according to key (Right). Dotted line shows the border of V1 (area 17) and V2 (area 18). (B) Blood vessel map (at 546 nm) corresponding to A. EGFP-expressing virus was injected into the boxed area, and a chronic transparent window was implanted above the exposed cortex for repeated optical and two-photon imaging. (C) Magnified image of the boxed area from A and B shows EGFP-labeled cells and their processes imaged by using two-photon microscopy (low digital zoom, magnification of 1×). The region and its processes can be precisely aligned with the functional OD map, based on the blood vessel pattern in B. (D) Magnified image of the boxed area in C shows labeled dendrites and axons. (E) Magnified image of the boxed area in D shows dendritic spines. This image corresponds approximately to a single pixel in A and B, allowing dendritic spine structure to be correlated with functional eye-specific drive on a pixel-by-pixel basis. Individual spines were outlined to measure the fluorescence of all pixels, and only the spines with an average fluorescence 1 SE higher than the neighboring neuropil (stippled circles in E) were counted as existing spines. (F) Quantification of fluorescent intensity (arbitrary units) for the three labeled spines and neighboring neuropil areas in E. Stippled line shows the detection threshold. Notice that area c falls below the detection threshold and was not counted as a spine. (Scale bars: A and B, 1 mm; C, 100 μm; D, 20 μm; E, 5 μm.)

Rapid OD shift evoked by MD and recovery in ferret V1. (A) Changes in deprived eye response (D), nondeprived eye response (ND), and ODI maps, measured by repetitive intrinsic signal optical imaging of the same region of V1 within a span of 9 d. Following the control image, the eyelid of the contralateral eye was sutured first for 3 h, opened briefly for a short-term MD data point (3 h MD), sutured immediately for 7 d, then opened for a long-term MD data point (7D MD), and left open for 2 d to enable successive recovery period data points [3 h, 1 d (1D), and 2 d (2D) recovery]. (Scale bar: 1 mm.) Dotted line shows the border of area 17 and 18. (B) Histograms of the ODI of all of the pixels in the map. Higher values represent nondeprived eye-dominated responses. (C) Mean deprived (blue) and nondeprived (red) eye responsiveness and ODI (black) at different time points. The values of all of the cortical pixels in the map were averaged for each corresponding time point. It is interesting to note that the MD effect is not linear across time: the initial 3 h MD evoked a similar ODI shift as the 3 h to 7 d MD. (D) Rapid OD shifts in four ferrets, evoked by 3 to 6 h MD and im-mediately followed by 3 to 6 h of recovery from MD.

A global shift of OD is evoked by brief MD. (A and B) OD maps of the same visual cortex before (A) and after 6 h of MD (B). Dotted line shows the border of area 17 and 18. (C) Pixel-to-pixel correlation of ODI before and after MD. Blue, four individual ferrets; red, average of four animals. The entire set of pixels was divided into 100 bins according to their ODI percentile before MD; the same group of pixels in each bin was then averaged after MD. Notice the ODI shift toward the nondeprived eye in monocular and binocular regions. (D) Greater loss of absolute responsiveness in deprived eye-dominated regions. The differential map (comparing pre-MD control and after MD) resembles the control map pattern. (E) Graph shows proportional change in responsiveness for the deprived and nondeprived eyes. The x axis shows the degree of pre-MD response from either eye across the pixels of the map; pixels can vary from near 0% to near 100% response amplitude from each eye. The y axis shows the shift in relative response after MD; note that both the deprived (“D”) and nondeprived (“ND”) lines are nearly horizontal, indicating that the percent response change is independent of the initial responsiveness of pixels to one or the other eye. Thus, although the proportional change is uniform, deprived eye-dominated regions show greater absolute loss of responses. (Scale bars: 1 mm.)

An activity-dependent model simulates experimental OD changes. (A–D) OD maps before MD (A), after MD (B), and simulated after MD by using an activity-dependent model with a global shift (C) and a competitive model (D). Dotted line shows the border of area 17 and 18. Simulation was applied to the deprived and nondeprived eye response maps before MD. In the activity-dependent model, the response amplitudes of all of the pixels in the deprived eye map were reduced by 40%, whereas all pixels in the nondeprived eye map were increased by 10%. The simulated OD map is the differential map between these two shifted maps. In the competitive model, pixels were altered by 25% based on the binocularity index (SI Materials and Methods). (E) Close correlation between the experimental and activity-dependent simulated OD maps after MD but not the competitive simulated OD maps. All the pixels of the control OD map were binned into 100 groups according to their percentile ODI, and the means of the same group of pixels were calculated in experimentally shifted OD map after MD and simulated OD maps.

Rapid remodeling of dendritic spines after short-term MD and recovery from MD. Spine turnover in the primary visual cortex of (A) a control ferret and (B) a ferret with MD. Red, intact spines; yellow, lost spines; green, new spines (based on comparison with the initial imaging session). Note that spines 23 and 19 in B reappear after the recovery period, whereas other spines (spines 26 and 27) are made at new locations at this imaging time point. (C) Short-term MD induces a significant decrease in spine number, which reverses after the eye is reopened (blue line). No changes in spine number were observed in control animals (pink line). (D) MD and recovery periods affect the turnover of spines. Spine loss and gain were increased by MD and recovery. Blue bars represent MD and pink bars represent control (*P < 0.05 vs. control).

Dendritic spine remodeling correlates with functional changes in defined OD regions within V1. (A) Spine turnover in four regions (marked 1–4 with squares in OD map, Center) with different OD preferences after 6 h of MD. Green arrows, new spines; red arrows, lost spines. (B-D) Rates of change of total number (B), loss (C), and gain (D) of spines as a function of ODI at each of 29 sites. Each site includes the analysis of all spines in an 80 × 60-μm region, and corresponds to a region of 6 × 4 pixels in the intrinsic signal imaging map (SI Materials and Methods). ODI was less than −0.001 at four sites with 436 spines; greater than −0.001 and less than 0 at 15 sites with 1,155 spines, and between 0 and 0.001 at 10 sites with 1,015 spines.