Functional Magnetic Resonance Imaging (fMRI) of the Visual Cortex with Wide-View Retinotopic Stimulation

Summary

We have developed techniques for mapping the visual cortex function utilizing more of the visual field than is commonly used. This approach has the potential to enhance the evaluation of vision disorders and eye diseases.

Abstract

High-resolution retinotopic blood oxygenation level-dependent (BOLD) functional magnetic resonance imaging (fMRI) with a wide-view presentation can be used to functionally map the peripheral and central visual cortex. This method for measuring functional changes of the visual brain allows for functional mapping of the occipital lobe, stimulating >100° (±50°) or more of the visual field, compared to standard fMRI visual presentation setups which usually cover <30° of the visual field. A simple wide-view stimulation system for BOLD fMRI can be set up using common MR-compatible projectors by placing a large mirror or screen close to the subject's face and using only the posterior half of a standard head coil to provide a wide-viewing angle without obstructing their vision. The wide-view retinotopic fMRI map can then be imaged using various retinotopic stimulation paradigms, and the data can be analyzed to determine the functional activity of visual cortical regions corresponding to central and peripheral vision. This method provides a practical, easy-to-implement visual presentation system that can be used to evaluate changes in the peripheral and central visual cortex due to eye diseases such as glaucoma and the vision loss that may accompany them.

Introduction

Functional magnetic resonance imaging (fMRI) is a valuable method to assess changes in regional neurovascular function within the visual cortex in response to stimuli, as changes in regional blood flow correlate to the activation of brain regions^1,2. High-resolution retinotopic blood oxygenation level-dependent (BOLD) signal measurements represent changes in deoxyhemoglobin, which are driven by localized changes in blood flow and blood oxygenation within the brain^1,2. BOLD activity patterns collected from fMRI data can be used to functionally map the peripheral and central visual cortex, as well as detect changes in the retinotopic map in response to visual impairment and neurodegeneration³.

Most previous fMRI studies made use of narrow-view (around ±12° of the central visual field) non-retinotopic stimuli or simple retinotopic stimuli with narrow-view visual stimuli, which provided limited functional parcellation of the retinotopic representation in the visual cortex and limited assessment to only the central visual field, excluding the periphery³. Consequently, narrow-view fMRI data has reported inconsistent BOLD percent changes in glaucoma patients^4,5,6. There is therefore a need for improved fMRI approaches to assessing the peripheral and central visual field, particularly in the evaluation of diseases such as glaucoma.

Glaucoma is the leading cause of irreversible blindness, affecting 10% of people by the age of 80⁷. Glaucoma is caused by the progressive, irreversible neurodegeneration of retinal ganglion cells, which are responsible for transmitting visual stimuli to the brain through the optic nerve. In primary open-angle glaucoma (POAG), the most common form of glaucoma, increased intraocular pressure causes thinning of the retinal nerve fiber layer (RNFL), leading to the loss of peripheral vision followed by peripheral and central blindess^8,9,10,11. Histological evidence from animal studies suggests that glaucoma additionally results in progressive neurodegeneration of the optic nerve, optic tract, lateral geniculate nucleus, optic radiation, and visual cortex^12,13. MRI technology offers a minimally invasive method of assessing both blood oxygenation and neurodegeneration in the visual cortex. In patients with glaucoma, MRI has found evidence of gray-matter atrophy in the visual pathway^13,14,15,16 and abnormal white matter in the optic chiasm, optic tract, and optic radiation^1,17,18.

To further explore the effects on visual processing, fMRI can be used to detect brain function in response to visual cues. The protocol herein describes a novel method to obtain a low-cost, wide-view retinotopic map using high-resolution retinotopy fMRI with wide-field (>100°) stimuli, as described by Zhou et al³. Visual stimuli of expanding rings and rotating wedges were used to elicit retinotopic mapping of the eccentricity and polar angle for fMRI. BOLD fMRI percent changes were analyzed as a function of eccentricity to evaluate brain function, corresponding to both central and peripheral vision. The BOLD fMRI percent change may be used to visualize activation throughout the visual cortex. These fMRI measures provide a reliable new method to evaluate neurodegenerative changes and their functional effects on the visual cortex found in eye diseases involving visual field defects, such as glaucoma.

Protocol

Research with human participants was performed in compliance with institutional guidelines at the University of Texas Health Science Center and Stony Brook University, with informed consent obtained from participants for these studies and use of their data.

1. Setup of MRI scanner and imaging protocols

1. For fMRI, use a 3T MRI scanner with multi-channel receiver head coils. Different field strengths can also be used but may present difficulties with signal-to-noise ratio (SNR) or distortion artifacts, so adjust accordingly. Use only the posterior half of the head coil for fMRI to allow for a larger viewing angle unobstructed by the anterior half of the coil.
2. Set up a T1-weighted magnetization-prepared rapid acquisition gradient echo (MP-RAGE) sequence with a repetition time (TR) of 2.2 s, echo time (TE) of 2.8 ms, field of view (FOV) of 176 mm x 256 mm x 208 mm, spatial resolution of 1 mm x 1 mm x 1 mm, bandwidth of 190 Hz/pixel, flip angle of 13°, and a scan duration of 3.1 min³.
3. Set up a gradient-echo, echo-planar imaging (EPI) sequence with a TR of 2 s, TE of 30 ms, FOV of 220 mm x 220 mm, in-plane resolution of 1.7 mm x 1.7 mm, 29 slices with a thickness of 3 mm, and a bandwidth of 1,500 Hz/pixel³.
4. Measure the dimensions of the head coil and scanner bore, and then construct a simple frame by cutting a polyvinyl chloride (PVC) pipe into suitable lengths and connecting them with PVC elbows. Obtain a mirror that is at least 25 cm wide and 15 cm tall and attach it to a plastic rod with screws (small holes can be drilled into the mirror).
   1. Attach the ends of the plastic rod to the PVC frame with nylon screws (Figure 1A). Make sure that the nylon screws are slightly loose to allow the mirror to be rotated by hand to optimize the angle for each participant.
5. Make a screen to go inside the MRI bore. Cut a segment of a rear-projection screen that is approximately the size of the MRI bore. Construct a frame that is the size of the bore and attach the screen to the frame with screws. Place the screen inside the scanner just behind the head coil to minimize the distance between the screen and mirror and maximize the FOV.
   NOTE: If the scanner bore is large enough, a single screen can be used for the participant to view directly instead of the mirror and rear projection screen setup. A projection screen attached to a thin sheet of wood for backing or a sheet of thin matte white plastic can be used as a screen and placed on the frame instead of a mirror. The projector should then be positioned and focused, such that it fills the screen and is in focus.

2. Participant preparation

1. Inform the participant about the procedure, risks, and benefits of the fMRI scan. Obtain their informed consent.
2. Ensure that the participant does not have any contraindications to MRI. This includes screening for pacemakers, metal implants, or claustrophobia. If you have any uncertainty, consult with a qualified radiologist or researcher, and exclude the participant from the study if any uncertainty remains.
3. Explain the visual stimulation protocol and the need for the participants to fixate on the central cross during the fMRI scans. Show the participant a short demonstration of the visual stimulation for instructional purposes to familiarize them with the procedure.
4. Carefully position the participant on the table of the MRI scanner to ensure that they are comfortable and relaxed. Provide earplugs and/or a sound-dampening headset to reduce the acoustic noise the participant will hear to protect their hearing.
5. Immobilize the participant's head in the posterior half of the head coil array, using foam padding on the sides of the head to ensure that the head is properly immobilized to reduce motion artifacts. Use the scanner's positioning system and move the table into the scanner bore.
6. Place the wide-view screen or mirror 10 cm from the patient's eyes (Figure 1B). Place the bore-sized screen from the back of the scanner bore just behind the head coil. Adjust the position and angle of the mirror/screen for each participant to achieve a consistent viewing angle.
7. Ensure that the participant is comfortable throughout the scan via communication through the intercom.

3. fMRI scanning of participant

1. Run a localizer scan with three orthogonal planes and scanner adjustments and calibrations for frequency adjustment and shimming.
2. Run an MP-RAGE anatomical scan to help position the EPI slices.
3. Create visual stimuli, as described in the following steps, using a program for running behavioral or psychological experiments.
4. At the start of the fMRI protocol, instruct the participant to fixate on the white cross (3° x 3°), which should be on top of a gray background at the center of the stimuli for 10 s.
   NOTE: The white cross will be shown before and after each visual stimulation paradigm for 10 s. Thus, the total fMRI stimulation test for each paradigm is 200 s.
5. Present the first visual stimulation paradigm (a series of rotating wedges) for a period of 30 s (giving an angular velocity of 6°/s) and cycle through six periods. The wedge stimuli should include 12 frames of rotating wedges (one scan with clockwise rotation and one with counterclockwise), extending to the edge of the screen/mirror (>100° visual field), with an 8 Hz contrast-reversing black and white (100% contrast) checkerboard pattern (Figure 2A).
6. Present the white cross once again for 10 s.
7. Repeat steps 3.4-3.6 with the second visual stimulation paradigm (a series of expanding and contracting rings) for a period of 30 s (expanding or contracting at 1.8°/s of the visual field) and cycle through six periods. The ring stimuli should include eight frames of expanding or contracting rings (>100° visual field), with an 8 Hz contrast-reversing black and white (100% contrast) checkerboard pattern (Figure 2B).
8. After completing the fMRI, move the table out of the scanner bore while instructing the participant to remain still. Remove the mirror/screen, place the anterior portion of the head coil in addition to the posterior, and move the table back into the center of the scanner.
9. Acquire a quick localizer scan in case of any movement and acquire an MP-RAGE scan with the full head coil.
   NOTE: An anatomical image with the whole head coil is needed for accurate registration for group analyses and reconstruction purposes.

4. Analysis of retinotopic fMRI data

1. Download and install the FreeSurfer application for MRI analysis (https://surfer.nmr.mgh.harvard.edu)²⁰.
   NOTE: FreeSurfer version 5.3.0 was used herein.
2. Obtain images in Digital Imaging and Communications in Medicine (DICOM) format from the MRI scanner. Convert the DICOM files to nifti format using the dcm2niix application (https://www.nitrc.org/projects/mricrogl)²¹.
3. Process the T1-weighted scan to provide a cortical surface reference, as described in the following two steps. Use FreeSurfer to convert structural data from nifti format to .mgz format (mri_convert command).
4. Use the recon-all command in a shell environment to perform automated segmentation and cortical reconstruction of the structural data.
   NOTE: This step can take over 20 h to complete.
5. Use the graphical user interface tksurfer to view the inflated hemisphere and virtually cut the visual cortex along the calcarine fissure, and select the occipital lobe. Use the mris_flatten command to flatten the visual cortex patch. Repeat this step for both hemispheres.
6. For the fMRI data, first remove the rest periods, with only the fixation cross presented, from the start and end of the data. Screen the fMRI data for artifacts or large movements.
7. Preprocess the functional data for spatial smoothing and motion correction. Model the retinotopic stimulus paradigm and apply a canonical hemodynamic response function to construct the response function.
8. Perform retinotopic phase-encoded analysis of the fMRI data using the FreeSurfer functional analysis stream (mkanalysis-sess, selxavg3-sess, and fieldsign-sess commands) to correlate the BOLD fMRI time series with a modeled response function and obtain phase-encoded retinotopic maps, with a significance level of p < 0.01 (Figure 3).
9. Visualize the results of the retinotopic maps with color-coded activation maps overlaid on the virtually flattened visual cortex using the tksurfer-sess command, and display using the rtview command.
10. Use the phase-encoded retinotopic maps from the wedge stimuli to help define the boundaries of the primary visual cortex (V1) and other extra striate areas (V2 and V3) by field sign maps (Figure 3A), along with anatomical landmarks and FreeSurfer atlases.
11. To calculate the BOLD response at different eccentricities, first use FSL Feat (http://fsl.fmrib.ox.ac.uk/fsl) to calculate statistical maps using a general linear model for each size of ring stimuli with a z-score threshold of Z > 2.3^22,23. If group analysis is being performed, calculate the second level analysis for statistical maps of group differences with FSL Feat to help determine the BOLD response at different eccentricities.
12. Co-register the fMRI images onto the reconstructed cortical surface using the FreeSurfer bbregister and tkregister2 commands to align the participant's fMRI data to the anatomical structural image of their brain and ensure accurate spatial alignment.
13. Group the ring stimuli by eccentricity for each of the eight frames. Manually draw regions of interest for different eccentricities based on the activated voxel regions for each frame. Take the BOLD percent changes and plot them as a function of eccentricity. Also, bin the eccentricity data into central (< ±12°) and peripheral (> ±12°) regions, where a ±12° visual stimulus is typical for retinotopic fMRI studies.

Results

Nine participants diagnosed with POAG (four males, 36-74 years old) and nine age-matched healthy volunteers (six males, 53-65) were evaluated using the aforementioned wide-view fMRI protocol, as previously described by Zhou et al³. POAG was confirmed clinically in patients with an open angle by assessment of the presentation of visual field defects consistent with glaucoma, optic disc cupping, and/or an intraocular pressure (IOP) greater than 21 mmHg³. A wide-view visual pr...

Discussion

The above protocol for utilization of wide-view retinotopic fMRI is an innovative method to evaluate the effects of vision loss and eye diseases on the brain. Through wide-field retinotopic mapping of the visual cortex with the use of a wider-view screen, this approach allows for a more comprehensive understanding of the visual system's functional organization. This could lead to a better understanding of abnormalities in the brain's visual processing system, which occurs in neurodegeneration, such as in glaucoma

Materials

Name	Company	Catalog Number	Comments
1/4"-20 nylon machine screws, knurled head thumb screw			to attach rod to PVC frame
1-1/4 inch PVC pipe			length of ~5-10 ft is needed
3T MRI scanner	Siemens
6-32 nylon machine screws, rounded head			to attach mirror/screen to rod
8-channel head array coil	Siemens
90 degree PVC elbow, 1-1/4 inch fitting
Acrylic mirror			Width and length of 25-30cm
Acrylic rod			1 inch width, ~ 2 ft long depening on size of scanner bore and head coil
E-Prime	Psychology Software Tools		to prepare and present visual stimuli paradigms
Plywood sheet, 1/2 inch thick			Size should be at least as large as the scanner bore. Cut as bore-sized frame for the projection screen
Rear projection screen			Size should be at least as large as the scanner bore