Visual Tract Integrity Before and After Gene Therapy in Congenital Achromatopsia

Abstract

Purpose: CNGA3 achromatopsia is a rare hereditary syndrome caused by dysfunction of cone photoreceptors. Visual information is therefore obtained only by rod photoreceptors, resulting in low acuity, photoaversion, and color blindness. Trials using gene therapy have been initiated recently, in which clinical improvement was subtle.

Methods: To explain this suboptimal outcome, we used diffusion tensor imaging to assess visual pathway integrity in 3 CNGA3 achromatopsia patients before and after gene therapy, and compared them with 16 normally sighted adults.

Results: No significant differences from normal subjects in optic tract and radiation were detected. Fiber integrity reduction was observed in the occipitocallosal fibers. These differences showed some normalization after treatment, but intersubject variability was evident. Specifically, the observed changes were related to radial diffusivities, reflecting fiber myelination or glial cell alterations.

Conclusions: Despite the fundamental role of cone photoreceptors in human sight, primary visual pathways in patients are comparable with those of healthy individuals and thereby fiber integrity is probably not an obstacle for recovery. Preliminary results suggest that the splenial fibers are less cohesive in naïve patients and regain some integrity after treatment. These findings add to previous reports on this rare population and suggest that novel information is processed within the visual cortex after treatment.

Translational Relevance: Patients with complete color blindness were treated using a novel gene augmentation therapy. Unfortunately, the patients did not experience a sudden eureka moment of being able to perceive the full spectrum of colors. In this study, we rule out fiber disintegration as the cause of their limited recovery.

Introduction

Congenital achromatopsia, also referred to as rod monochromacy, is a rare, autosomal recessive visual disorder causing dysfunction of the cone photoreceptors in the human retina. Because cones are responsible for color vision, day vision, and high-resolution foveal sight, patients suffer from complete color blindness, photoaversion, low visual acuity, and central scotomas.¹ 

CNGA3 achromatopsia does not result in the absence of cone photoreceptors, but in nonfunctional ones. The CNGA3 gene encodes a specific subunit of the cGMP-gated cation channel, located in the cone's receptor membrane, and it is crucial for the phototransduction cascade.² In CNGA3 achromatopsia, the photoreceptors remain viable but are not activated when exposed to light. 

The fact that cone photoreceptors remain intact and potentially functional, along with the highly specific and monogenic nature of the disorder, makes CNGA3 achromatopsia a suitable candidate for gene therapy. Promising results were demonstrated in ovine models of the disease, with both behavioral improvement and proven cone functionality in electroretinography tests.³ The current understanding is that expressing the CNGA3 genes in human retinas will allow for reactivation of patients’ cones and hopefully restore visual function. In recent years, a few early stage human trials examining the safety and efficacy of the gene therapy have been commenced. In those trials, a viral vector carrying the missing cone-specific opsin promoter is being subretinally injected into the patient's eye. 

To date, there are only a handful of papers reporting patients’ post-treatment visual functions and cortical activity. Examining two patients using functional magnetic resonance imaging, we have shown a decrease in population receptive field (PRF) sizes in early visual cortex after surgery (although still greater than those of the normal population), a measurement that may hint at a potential improvement of visual resolution. Similar results of partial improvement in cortical visual maps were obtained in children who underwent gene augmentation therapy.⁴ Behaviorally, testing four patients, we have shown a mild improvement in their color perception⁵ as did another trial in one patient.⁶ Although a mild improvement is seen in certain aspects of cortical activity and color percept, patients’ values are still far from the normally sighted population and none of the centers reported improvement in acuity. This finding suggests that, although the visual cortex responds to restorative procedures, the response is somewhat limited. 

When considering the possible factors limiting clinical recovery, it is helpful to think of them as existing on any one of three levels: failure in the retina itself, possibly owing to inadequate cone reactivation; failure of information transmission from the retina to the cortex along the afferent visual pathways; and failure of cortical processing of novel information. Here, we specifically explore the information transmission level by studying the integrity of patients' visual fibers. We perform diffusion tensor imaging (DTI) analysis of the visual tracts to reveal whether congenital achromatopsia causes alterations to the white matter microstructure and whether these changes respond to restorative surgery or alternatively serve as an obstacle to recovery. 

Methods

Patients

Three patients with confirmed CNGA3 achromatopsia were recruited at the Hadassah Medical Center in Jerusalem, Israel, as part of a single-site study. The same subjects are enrolled in a multicenter trial studying the efficacy and safety of gene augmentation for CNGA3 achromatopsia (ClinicalTrials.gov ID NCT02935517). Briefly, patients were treated with a nonreplicating, rep/cap-deleted, recombinant adeno-associated virus vector expressing the CNGA3 gene under control of an engineered PR1.7 cone opsin promoter. The viral suspension was injected into the subretinal space adjacent to the macula of each patient's worse-seeing eye. 

Patients completed magnetic resonance imaing scans before and after treatment. Patient 1 was scanned before and 4 months after the procedure, patient 2 was scanned before and 7 months after the procedure, and patient 3 was scanned before and 4, 8, and 12 months after the procedure. Similar scans were acquired for 16 normally sighted control participants. Three of the control participants were scanned a second time 3 to 4 years later to demonstrate consistency of the measurements. The measurements of these three patients at the initial scan and at the follow-up scan were compared with those of the rest of the control group using the same tests used for the patient group before and after treatment. 

The study received ethical approval from the Committee on Research Involving Human Subjects of the Hebrew University-Hadassah Medical School (Helsinki Committee of Hadassah Medical Organization) and the Israel Ministry of Health. All research was performed according to the guideline and regulation specified by the ethical Helsinki committee of Hadassah Medical Organization and as listed in the Declaration of Helsinki. Patients gave written informed consent for both the initial trial and follow-up imaging analysis. 

Data Acquisition

Scans were performed in the ELSC brain center in the Hebrew University, using a 32-channel coil in a 3T MAGNETOM Skyra scanner (Siemens Healthcare, Erlangen, Germany). Anatomical data were obtained using an MPRAGE sequence (TR, 2300 ms; TE, 2.98 ms; flip angle, 9°; isotropic voxel size, 1 mm; 160 axial 256*256 mm slices). DTI data were obtained with a diffusion sequence (TR, 4000 ms; TE, 96.2 ms; flip angle, 90°; slice thickness, 1.5 mm; matrix, 128 × 128; inplane resolution, 1.5 × 1.5 mm; B0, 2000; b-value gradients were applied along 64 different diffusion directions, and 1 average). To create eccentricity maps that were used for delineating cortical foveal and parafoveal areas, functional data were obtained using an echo planar imaging sequence (TR, 1000 ms; TE, 34.4 ms; flip angle, 62°; isotropic voxel size, 2.5 mm; 48 slices). A 2° wide bar, consisting of a black (0.37 cd/m²) and white (180 cd/m²) checkerboard pattern (1° black and white squares) flickering at 2 Hz, moved parallel to its orientation in 1°/s steps, taking 16 steps across the 16° screen. The movement was performed eight times; each time, the bar wiped the screen in a different direction (horizontally left/right, vertically up/down, in four diagonal directions). Overall, the stimulus covered 16° of visual angle (some of these data were included in McKyton et al.,⁷ for further details see there). 

Data Preprocessing

Fiber acquisition and analysis was similar to those previous published by our group.⁸ We describe them here briefly. DTI processing and analyses were performed using the mrVista open-source package for Matlab (http://vistalab.stanford.edu/software). Preprocessing included correction of eddy distortions and head motion. Diffusion data were matched to a T1-weighted anatomical volume which was aligned, as standard, to a line connecting the midpoints of the anterior commissure and the posterior commissure (the aligned volume henceforth referred to as the AC–PC space). Retinotopic mapping analysis was performed using BrainVoyager software (Brain Innovation, Maastricht, the Netherlands). Functional scans were preprocessed by correcting slice scan time and three-dimensional motion and by filtering out temporal frequencies. Functional scans were aligned to the anatomical scans, after which both anatomical and functional scans were realigned to match Montreal Neurological Institute space coordinates. 

Fiber Delineation and DTI Analysis: Optic Tract

The optic chiasm and both lateral geniculate nuclei (LGN) were identified manually using anatomical landmarks on the T1 map.⁹ Regions of interest (ROIs) corresponding to the beginnings of the left and right optic tracts and both LGNs were positioned on T1 maps of each subject. Using the ConTrack probabilistic tractography algorithm,¹⁰ the most likely pathway between ROIs in each hemisphere was calculated. A set of 10,000 potential pathways was generated, and the 1000 highest-scoring fibers were chosen, representing the optic tract. 

Optic Radiation

An anatomical estimate of the calcarine sulcus in each hemisphere was selected as an ROI using the high-resolution T1 images, as previously described by Sherbondy et al.⁹ The optic radiation (OR) in each hemisphere was estimated as the most likely pathway between the LGN and the calcarine ROI. The top 15,000 out of 75,000 potential pathways were selected. 

Foveal and Parafoveal Fibers

We used PRF analysis to divide primary visual regions into fovea and parafovea. From voxels within the early visual cortex (defined by anatomical atlas:¹¹), voxels fitting the PRF model with an R of greater than 0.34 were used. Eccentricities of 0° to 3° were defined as foveal ROI, and eccentricities of 5° to 8° were defined as parafoveal ROI.¹² Fibers of the OR terminating in either (left or right) foveal ROIs or parafoveal ROIs were selected and parceled from the entire OR. Because the functional data enabling construction of retinotopic maps was available in 13 out of 16 controls, the control group for this analysis includes only 13 controls. 

Callosal Fibers

Callosal segmentation was performed as described by Huang et al.¹³ The corpus callosum was segmented based on the estimated cortical projection zone of posterior callosal fibers. The procedure was initiated by selecting the corpus callosum as a ROI and using a deterministic streamlines tracing technique algorithm to trace all fibers crossing it. To define the occipital lobe, an ROI was selected manually in the coronal plane posterior to each parieto-occipital sulcus, according to the abovementioned method of Huang et al. The subsets of fibers that pass through the corpus callosum and each of these ROIs were identified. The cross-sectional area of these fibers in the plane of the corpus callosum was calculated, and the cross-sectional area of the whole corpus callosum was also estimated. All ROIs were identified by a trained medical student (H.A.). 

In all pathways, clearly misidentified fibers were manually removed using the Quench component for visualization and segmentation of tractography results. Identified fiber pathways were sampled using mrVISTA for DTI parameters, namely, fractional anisotropy (FA), axial diffusivity (AD), and radial diffusivity (RD). Parameters were calculated by fitting a diffusion tensor model to each voxel. To avoid partial voluming with non-white matter, diffusion properties were estimated by combining data in a weighted fashion around the dense core of the fibers.¹⁴ Diffusion measures along the tract were resampled at 30 positions along the optic tract (from chiasma to the LGN), at 50 positions along the OR (from the LGN to the calcarine sulcus), and at 50 positions along the splenial fibers (from the corpus callosum to the occipital lobe), calculating the FA, AD, and RD at each of these nodes. This process creates an anatomically independent representation of the fiber properties and thereby supports comparing data across subjects. Additionally, 5 points were removed from the beginning and end of the fibers, where partial volumes are especially prevalent. 

Data analysis and fiber presentation (in the figures) were done using the mrVista and AFQ packages.¹⁴ 

Statistical Methods

Statistical analysis was performed using SPSS. Control data is presented as mean (±SD between controls). For each fiber, analysis was performed for the average value of the entire fiber and for each location along the fiber. To correct for multiple comparisons, a familywise error (FWE)-corrected alpha for pointwise comparisons and a FWE-corrected cluster size were calculated using the Mann–Whitney test for each fiber at a P value of less than 0.005.¹⁵ A between-group effect along the fiber was considered significant if it passed any of these criteria. All tests were two tailed, and statistical significance was determined at a P value of less than 0.005. 

Results

Cohort Characteristics

Three achromatopsia patients (one female; aged 27, 30, and 31 years) participated in the study. Sixteen normally sighted control participants (11 females; mean age, 27.3 ± 8.4 years) were enrolled to serve as a control group. 

Afferent Visual Pathway; Optic Tracts and Radiations

Optic tracts and radiations were delineated on a participant-by-participant basis (Fig. 1A). Diffusivity measurements were analyzed along the fibers at the different resampling points, enabling comparison along the tracts. No differences in the mean FA, RD, and AD values were found between patients and controls (either each side alone or when averaged). Similarly, when comparing each location along the afferent fibers, no differences between patients and controls were detected. FA results throughout the fibers are presented in Figure 1B. The mean values of all three diffusivity parameters and P values comparing subjects to controls are presented in Table. 

Figure 1.

 

Afferent visual pathway integrity. (A) Visualization of the optic tracts (blue) and optic radiations (ORs) (red) of a representative control, as delineated via fiber tractography, superimposed on an axial cortical T1-weighted view. (B) FA profiles along the optic tracts (left) and ORs (right). No significant difference between patients and controls was found. OC, optic chiasm; VC, visual cortex. Gray area represents standard deviation for the control values. Values are averages of two hemispheres.

Table.

 

Average DTI Parameters for Visual Pathways (± Standard Deviation Along Each Fiber)

 

Additionally, no significant differences were found when comparing fiber parameters in patients before treatment with these parameters in patients after treatment. 

Afferent Visual Pathway: Optic Radiation Foveal Fibers

The visual pathways transfer information from the entire retina, i.e. input received from both rods and cones. To delineate the effect of cone input (or lack thereof), a subgroup of the OR fibers representing the fovea was analyzed separately. These fibers are specifically wired to foveal early visual areas, relaying information originating in the cone-dense fovea.¹² 

Diffusivity measurements were again analyzed along the ORs at the different resampling points. No significant differences in the mean FA, RD, or AD values were found between the foveal and the parafoveal fiber subgroups, or when comparing patients with controls in either of the fiber subgroups both before and after treatment (although a trend was apparent in foveal fibers before treatment). A larger group of patients may allow to reveal differences in this fiber as well (Table; Fig. 2). 

Figure 2.

 

OR foveal and parafoveal fiber integrity. (A) Visualization of the foveal (pink), parafoveal (orange), and the entire (red) OR fibers of a representative control, as delineated via fiber tractography, superimposed on an axial cortical T1-weighted view. (B) FA profiles along the foveal (left) and parafoveal (right) ORs fascicles. VC, visual cortex. Gray area represents standard deviation for the controls’ foveal fibers’ FA values.

Occipito-occipital Pathways

Splenial fibers, which originate in the occipital lobes and pass through the posterior corpus callosum, were segmented (Fig. 3A). Before treatment, mean splenial FA was significantly lower and mean RD was significantly higher in patients compared with controls (Table). Comparing subjects and controls along the length of the fiber showed a reduction in FA and an increase in RD along significant portions of the fiber (FA points 12–27 cluster size corrected, 12–23 FWE alpha corrected; RD points 8–30 cluster size corrected, 8, 12–22 FWE alpha corrected) (Fig. 3B). After treatment, the mean values of FA and RD were again lower and higher, respectively, in patients compared with controls. A point-by-point comparison along the fibers showed a reduced FA in points 13 to 22 with cluster size correction, 13 to 21 with FWE alpha size correction. The RD was increased in points 13 to 22 with cluster size correction and 14 to 21 with FWE alpha size correction. 

Figure 3.

 

Occipito-occipital (splenial) pathway integrity. (A) Visualization of the splenial fibers (green) of a representative control as delineated via fiber tractography, superimposed on an axial cortical T1-weighted view. (B) FA profiles along the splenial fibers for each patient. Gray area represents standard deviation for the controls’ FA values. Colored bar at bottom of figures; locations along fibers where patients and controls significantly differ with a P value of less than 0.005 FWE corrected clusters. (C) Scatter plot of the average splenial fibers’ radial and axial diffusivities for each patient (before and after treatment) and control. Each point shows the average of the right and left splenial tracts. The two standard deviation covariance ellipsoid of the normal population is shown. Filled and empty shapes are patients before and after treatment, respectively. Two of the patients show normalization after treatment. (D) Corpus callosum and splenial cross section in each patient (before and after treatment) and in the control group.

However, because the posttreatment differences between patients and controls seemed reduced compared with pretreatment results, we further studied the individual effects. Figure 3C displays a scatter plot of the average splenial fiber RD and AD for each patient and control. Each point shows the average of the right and left splenial tracts, which did not differ significantly. Before treatment, all three patients had a higher than expected RD relative to their AD, rendering them outside the two SD covariance ellipsoid of the normal population. After treatment, the RD for patients 1 and 3 normalized. Repeated posttreatment measurements were available for patient 3, and they demonstrated consistency in his posttreatment results (Supplementary Fig. S1A). 

Finally, the splenial fibers were used to parcel the corpus callosum (Fig. 3D). In the normal population, there is a correlation between the cross-sectional area of the splenial fibers and the cross-sectional area of the entire callosum.^16,17 In contrast, all patients had a small occipital-callosal fiber cross section and a normally sized corpus callosum (Mann–Whitney-derived P values were 0.361 for the corpus callosum cross section, 0.014 for the cross section of splenial fibers, and 0.008 for the ratio between them). These measurements did not change significantly following treatment (P values derived from the Wilcoxon signed-rank test were 0.285 for the corpus callosum and 0.593 for the splenial fibers). 

Because our cohort, although unique, is very small, we wanted to verify that these changes between scans result from treatment and not from variance between repeated scans. To that end, we performed an additional scan for three normally sighted subjects from the control group; all were scanned 3 or 4 years after their initial scan. No significant differences were found between this subgroup of controls and the rest of the control cohort in either of the scans (statistical results are presented in Supplementary Table S1). Comparison along the fibers' length is presented in Supplementary Fig. S1B, showing consistency of the integrity along the callosal in the controls as opposed to post-treatment alterations in the patients. 

Discussion

The distinct visual defect of CNGA3 achromatopsia patients provides a unique view of the visual system and may be seen as a controlled model for assessing the role of cone photoreceptors in visual development and function. Even more unique are the patients of the AGTC-402 trial, who underwent novel restorative treatment, allowing further examination of the potential for normalization of this congenital disease. 

Initial results show that clinical improvement is limited following gene therapy (see Anderson et al.⁶ for similar results in a different, larger cohort). This finding is somewhat expected, because visually impaired subjects do not necessarily recover after restorative treatment.¹⁸ The visual system requires visual input for healthy development and maintenance.^16,19 Long periods of visual impairment have been shown to affect its structure, causing reduced white matter in visual pathways and altered gray matter organization in the visual cortex and thalamus.^20,21 Thus, after a certain point in early life, the visual system loses most of its plasticity, and the lost functions cannot be restored with sensory recovery. It is difficult to establish to what extent this is the case in CNGA3 patients. The disorder limits perception of visual stimuli from birth, including during critical developmental stages. Yet it is important to note that these patients function independently in their daily visual activities, suggesting that a substantial amount of information is still received and processed through the unaffected rods and might be sufficient for relatively normal development of the visual system. 

The achromatic visual cortex, however, is clearly altered by the abnormal input. Baseler et al.²² demonstrated reorganization of the visual cortex in achromatopsia patients, causing a reduction in the size of areas representing foveal input. Treated patients' functional magnetic resonance imaging results also show cortical deficits in the form of low computational resolution which somewhat normalized following tratment.⁷ Recently, Anderson et al.⁶ reported the results of gene augmentation therapy in nine adult patients with another mutation causing congenital achromatopsia (CNGB3 mutation), describing some differences in retinotopic maps in patients compared with controls. These differences did not change after treatment in most patients, but did normalize to some extent in the one patient showing clinical improvement. 

When examining cortical morphology, Molz et al.²³ and Lowndes et al.²⁴ demonstrated a reduction in gray matter volume and cortical surface area in the visual cortex of achromatopsia patients. Increased thickness of the visual cortex, previously associated with total congenital blindness, was shown to be specifically localized to areas representing foveal input.^23,24 

The goal of the present study was to examine the involvement of white matter in the disorder and the potential of this involvement to inhibit recovery. Because the CNGA3 gene is not expressed outside the retina,² changes observed in fiber properties prior to treatment represent either irregular development of these fibers or a response to relatively sparse input. 

In our analysis, there was no significant difference in the fiber integrity of the afferent visual pathways (optic tracts and ORs) between patients and healthy controls. Furthermore, no change was observed in these fibers after treatment. Although it is well-established that the integrity of afferent visual fibers is reduced in congenital visual disorders,^16,25,26 this is usually demonstrated in impairments caused by direct nerve deafferentation, retinal damage or anatomical distortion. In cases without direct neuronal damage, the correlation is not as clear. For example, in amblyopia subjects, the mean diffusivity may be increased, but FA is within the normal range,²⁷ and in congenital achiasmia, the integrity of the OR remains unaltered despite anatomical abnormalities.²⁸ In this regard, our findings of preserved fiber integrity suggest that information from the still-functioning rods in the retina, despite being diminished, suffices for the development of normal optic tracts and ORs. 

Even when focusing our analysis on the OR foveal subgroup, which represents fibers that transfer information mainly from the cone photoreceptors, integrity seemed intact. It should be noted that this foveal subgroup of fibers contains rod-derived fascicles as well, which might mask some of the integrity change of cone-representing fibers. Synaptic plasticity within the retina may also contribute to the integrity of these fibers: after cone-specific receptor death or degeneration, bipolar and ganglion cells can reorganize their neural network.^29,30 Haverkamp et al.³¹ showed how cone–bipolar cells form ectopic synapses with rod receptors in a mouse model of CNGA3. It is reasonable to think that this process could occur in achromatopsia patients, causing foveal fascicles to receive information from rod receptors and retain normal development. 

The only disease-related white matter alteration that we found was reduced fiber integrity along the occipital–callosal fibers, as reflected through both mean value and point-by-point comparisons. This effect was found when comparing pretreatment patients to controls, but not when comparing the follow-up subgroup of controls to the rest of the group. The occipital–callosal fibers transfer information between each occipital lobe and the contralateral hemisphere and play a role in many visual functions. Works examining patients with lesions around the splenium or who underwent commissurotomies suggest that these fibers’ integrity is correlated with visual acuity, object identification and complex visual tasks such as reading.^17,32,33 In general, the fibers represent a slightly later phase of visual processing, after the initial cortical computation. In naïve patients (before treatment), the fibers' abnormal structure can be thought of as a downstream effect of the changes in the visual cortex, with the abnormal processing resulting in a less cohesive connection between occipital lobes. In other words, the achromatic visual system is not adapted for the complex visual functions facilitated by these fibers. Similarly, just as changes were seen in cortical mapping after treatment, integrity increased in the splenial fibers, promoting information transfer between occipital lobes and potentially enabling more complex visual tasks. A point-by-point comparison did show a significant decrease in fiber integrity that persisted after treatment, yet this decrease was not as striking as before, only reflected in small portions of the fiber. 

The splenial fibers' integrity alterations reflected through FA values were driven by RD values; relatively high pretreatment RD values were normalized in two patients after treatment. Change in the RD is considered reflective of fiber sheath myelination and the presence of support cells. Usually, an increase in the RD, especially when not accompanied by AD change, is correlated with a decrease in myelination and presence of fewer cell membranes.³⁴ Therefore, it is likely that our findings indicate that splenial fibers in CNGA3 achromatopsia patients differ from those of healthy individuals in the state of myelination or in the activity of myelinating cells. Changes in myelination and glial cell activity are among the mechanisms that facilitate plasticity in the adult brain.³⁵ Myelination is suggested to occur after learning novel skills, such as piano playing or reading,^36–38 and it is observed in animal models after motor training.³⁹ It is reasonable to think that, after cone reactivation, although behavioral changes are limited, more information is transferred through the splenial fibers, causing a process of increased myelination that is reflected in a decrease in the fibers’ RD. 

The occipital–callosal fibers of patients also have a smaller cross-section in the plane of the corpus callosum. Because theses fibers predominantly transfer visual information, their relatively small size indicates a callosum less optimized for visual information processing. This process may be caused by a lack of sufficient visual input during the development of the corpus callosum, as has been observed in congenitally blind individuals.^16,38,40 

It is important to note the limitations owing to the small sample size of the study and the interpatient variability. To that end, we used a relative rigorous statistical threshold (P < 0.005), and we scanned healthy subjects again following a time interval to demonstrate the stability of the fibers' microstructure parameters. Because CNGA3 achromatopsia is a rare disorder, and restorative treatment is only being used in initial trials, it is impossible to assess large groups of suitable subjects. Currently, statistical power greatly limits any analyses, and chance may play a considerable role in apparent findings. Therefore, results should be interpreted with caution, and the conclusions drawn from them should be taken merely as a starting point for more extensive future research. We suggest considering the current article a pilot to a larger, multicenter study. 

Conclusions

Our results suggest that recovery after gene therapy for CNGA3 achromatopsia is not limited by the transmission of cone receptor-derived information from the retina to the cortex. The relatively preserved cohesiveness of the afferent visual pathways means that these pathways are not a limiting factor for recovery, leaving hope for further recovery in the future. It is likely that the callosal fibers' integrity changes either reflect or help facilitate the cortical changes and subtle clinical improvement in patients following treatment. These findings, although preliminary, may help us to better understand the way that the visual cortex responds to abnormal input and potentially play a role in assessing patients for restorative treatment or follow-up. 

Acknowledgments

Supported by a grant from the Israel Science Foundation (ISF #2389/22), Jerusalem, Israel. This grant does not create any conflict of interest for any of the co-authors. 

Data Availability Statements: The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. 

Disclosure: H. Abramovitch, None; A.S. Bick, None; N. Guy, None; D. Elul, None; A. Mckyton, None; E. Banin, None; N. Levin, None