Abstract
Objectives: To investigate the pattern of functional organization in the human visual cortex through electrical cortical stimulation.
Methods: Electrical cortical stimulation was applied to the occipital cortex and adjacent cortices using subdural grid electrodes in 23 epilepsy patients. Diverse visual responses were recorded. These responses were divided into different categories according to the specific response modalities, such as form, color, and motion. Form visual responses were further subdivided into simple, intermediate, and complex responses. The cortical localization of subdural electrodes was identified using MRI-CT coregistration. The cortical distribution of different visual responses was projected into three-dimensional surface renderings of the brain. The distribution and frequency of subdural electrodes showing different visual responses were quantified by calculating the percentage of the number of electrodes showing one specific type of visual response at the corresponding anatomic region to the total number of electrodes in all brain regions that produced the same response.
Results: Simple form responses were obtained mostly at the occipital pole and the inferior occipital gyrus (47.4%) and the striate cortex (42.4%). Intermediate form responses occurred mainly on the peristriate cortex (52.5%) and the lateral occipital (28.0%) and fusiform gyri (19.5%). Complex forms were produced by stimulation of the basal temporo-occipital region (57.6%) and the lateral temporal or lateral temporo-occipital junctional region (42.4%). Color responses occurred on the basal occipital area, mostly at the fusiform (40.0%) and lingual gyri (36.0%). Moving sensations were evoked by stimulation of the basal temporo-occipital (28.4%) and the mesial parieto-occipital or temporo-parieto-occipital junctional regions (23.9%).
Conclusions: Different modalities of vision, such as form, color, and moving sensation, appeared to be distributed and organized in different areas of the human visual cortex.
Recent research suggests that various components of vision, such as form, color, and movement, are systematically organized in the perception of a visual image and that there are at least three or four different neural pathways for vision in the brain.1
The concept of functional organization or specialization in the visual cortex was not established until the middle of the 20th century. The studies that did so were performed on animal models. Unfortunately, there are limitations in applying these results to the humans, and the direct stimulation and recording of human visual cortex are not usually feasible. In previous studies, the analysis of evoked potentials after specific visual stimuli usually was performed on macaques or rhesus monkeys. Those studies suggested that there are many visual areas outside the striate cortex. One area (V5) is specific for visual motion; another area (V4) is specific for color.2-4 Clinical studies in patients with cortical lesions suggested that specific regions of the extrastriate or association visual cortex may be specialized for form, color, and motion. In patients with focal brain lesions, selective agnosias were reported for inanimate or animate objects or for familiar faces only (prosopagnosia). Some patients with localized damage to the basal temporal cortex showed loss of color vision only (achromatopsia) but had reasonably good visual acuity preserved.5-7 Conversely, bilateral damage to the medial temporal or medial superior temporal cortex may produce selective loss of movement perception.8,9 Recently, PET studies with colored and moving visual stimulation showed that the lingual and fusiform gyri were critically involved in the color vision, and the temporoparieto-occipital junction was related to the perception of movement.10
Cortical electrical stimulation recently has been used to localize areas of specific brain functions, and extraoperative cortical stimulation enables testing in greater detail and over a longer period of time. Recently a study of human visual-cortex stimulation demonstrated a direct relationship between specific cortical regions and corresponding visual fields.11 However, few studies have been performed in human brain using cortical stimulation, and the precise functional localization of the human visual cortex remains to be defined.
This study investigated functional organization of the human visual cortex for different visual attributes by direct cortical electrical stimulation. Electrical cortical stimulation was applied to the occipital cortices and adjacent areas of 23 patients with epilepsy, via subdural electrodes. The subdural electrodes were precisely localized on the cortical surface by the MRI-CT coregistration technique.
Methods.
Twenty-three patients with epilepsy who had subdural electrodes placed over their occipital cortices and adjacent cortical areas were included. Twelve patients were men, and 11 were women. The patients’ ages ranged from 16 to 41 years (mean, 26.7 years). All patients underwent presurgical evaluation including electrical cortical stimulation.
Electrical cortical stimulation.
Electrical cortical stimulation was performed with electrical current directly applied to the brain through metal electrodes using a Grass S12 Isolated Biphasic Stimulator (Quincy, MA). The electrical stimulation was carried out with a direct current of 300 μV of positive–negative square waves, a duration of 0.3 msec, and trains of 50 hertz for 5 seconds. The patients confronted a white board with vertical and horizontal meridian lines and a red-colored central fixation point. If a visual hallucination or illusion occurred, the patient was asked to draw the outline, shape, color, location, and motion of the response as precisely as possible on a white paper with the same vertical and horizontal meridians. The intensity of the electrical current was increased gradually from 1 mA until clinical response or afterdischarge was observed. If no response was observed using 15 mA, that cortical region was considered a functionally silent area. Clinical responses with afterdischarges were excluded from the data.12 If initial bipolar stimulation of two adjacent electrodes produced visual response, each electrode was stimulated separately with a reference electrode (that produced no subjective or objective response at stimulation) to identify the electrode responsible for the specific visual response.
Once diverse types of visual responses were obtained, they were divided into different categories according to the specific response modalities such as form, color, and motion. Additionally, responses with diverse forms were further subdivided into simple, intermediate, and complex responses. The simple form was defined as a white or black spot or a blob of flashing light. The intermediate form was defined as a geometric shape, triangle, diamond, or star. The complex form was defined as a formed visual hallucination (such as animals, faces, parts of the body, landscapes, or scenes from one’s memory) or illusion (such as waviness, distortion of vision, or visual field defect). The intermediate and complex forms were classified as black, white, or colored, and all forms were classified as still or moving.
Color response was determined only by the presence of color, without considering the shape or motion. Moving response was defined by perception of something moving, regardless of the shape or color. For example, the stimulation of one electrode produced a still, complex form with color, and another electrode showed a moving complex form without color.
Localization of subdural electrodes with MRI-CT coregistration.
Before surgery, an MRI was obtained with a GE Signa 1.5-tesla unit (GE Medical Systems, Milwaukee, WI). A SPGR (spoiled gradient recalled) three-dimensional volumetric MRI was acquired in 124 contiguous no-gap coronal slices with 1.5- to 1.8-mm thin sections. A repetition time of 33.33 seconds, an echo time of 7.0, and a 22-cm field of view were utilized. The sampling matrix was 256 × 192, and the data were reconstructed with a 256 × 256 matrix by means of sinc [sinc (x) = sin (x)/x] interpolation. A GE Hispeed Advantage CT scanner (GE Medical Systems) was used. The scanning parameters were 120 kVp (kilovolt [peak]), 220 mA, 1 second helical, 3 mm thickness without gap, algorithm of standard, display field of view of 21 cm, scan field of view of head, and a 512 × 512 sampling matrix.
Image-processing procedures were performed with Analyze 7.5 software (Biomedical Image Resource, Mayo Foundation, Rochester, MN) and the Sun Ultra 1 Creator workstation (Sun Microsystems, Mountain View, CA) in our neuroimaging laboratory. Image processing consisted of three steps: 1) MRI-CT registration and transformation; 2) subdural electrode of segmentation in CT images transformed to the MRI; and 3) multiobject rendering.
In MRI-CT registration and transformation, the cerebral area of the MRI was segmented by autotracing and a manual boundary-limit decision. The corresponding CT area was segmented by manual tracing. Before two images were registered, the brain structure on the MRI and CT was segmented and saved in 16-bit gray-scale and one-bit binary images. The one-bit CT image was transformed to MRI by a Chamfer surface-matching algorithm. This process made the most suitable transformation matrix. The mean matching error was less than 5 mm in diameter.
The second step was subdural electrode segmentation of transformed CT. By manual tracing, subdural electrodes were segmented from 16-bit transformed CT and saved in 16-bit gray scale and one-bit binary images.
Finally, multiobject rendering was done. Original MRI (voxel size, 0.86 × 0.86 × 1.7 mm) was reconstructed to isotropic voxel size (0.86 × 0.86 × 0.86 mm) to represent real brain size. The segmented subdural electrodes were also reconstructed to isotropic voxel size (0.86 × 0.86 × 0.86 mm). To overlay subdural electrodes on the cortical surface, the multiobject mode was used. The voxel-gradient shading-rendering with transparency was used.13,14
Analysis.
To analyze and compare the stimulation data among patients, all electrodes of individual patients were represented on a common composite brain map (figures 1 through 3⇓⇓). In the composite brain map, one square replaced one subdural electrode, and the width and length of squares were each 1 cm. Using the MRI-CT coregistration technique, the anatomic locations of all subdural electrodes producing visual responses were depicted by squares in the common composite map. As anatomic landmarks we used the collateral sulcus, calcarine, and parieto-occipital fissures, the posterior end of the occipital pole, and the lateral edge of the temporal lobe.
Figure 1. The lateral, midline sagittal, and basal views of the composite map of three-dimensional MRI surface rendering for 23 patients and the distribution of electrode positions showing responses of simple (A), intermediate (B), and complex (C) forms. One square represents the position of one subdural electrode on the cortical surface. The positive response rate of each visual response at one electrode location was calculated by dividing the number of patients showing one type of visual response at stimulation of one square location by the total number of patients having an electrode placed and stimulated at that square, and multiplying by 100%. Locations with positive rates less than 10% or at least 10% for each visual response were marked differently.
Figure 2. The lateral, midline sagittal, and basal views of the composite map of three-dimensional MRI surface rendering for 23 patients and the distribution of electrode positions showing responses of color (A) and moving sensations (B). One square represents the position of one subdural electrode on cortical surface. The positive response rate of each visual response was calculated by dividing the number of patients showing one type of visual response at stimulation of one square location by the total number of patients having an electrode placed and stimulated at that square, and multiplying by 100%. Locations with positive rates less than 10% or at least 10% for each visual response were marked differently.
Figure 3. Electrode positions producing scotoma or visual-field defect (A) and visual illusions (B), such as micropsia, macropsia, or metamorphopsia, during electrical cortical stimulation. One square represents the position of one subdural electrode on cortical surface.
The distribution of each type of visual response was expressed on the common composite brain map according to the positive response rate. The positive response rate of a specific visual response is the percentage of the total number of patients having an electrode placed and stimulated at a square who show one type of visual response to the stimulation of that square location. The positive response rate for a specific visual response at each square in the composite brain map was calculated by dividing the number of patients showing one type of visual response at stimulation of one square location by the total number of patients having an electrode placed and stimulated at that square, and multiplying by 100%. For example, if eight patients had electrodes placed at one square of a certain location and all of them were stimulated, the total number of patients was eight. If two of eight saw simple forms (tiny black spots) at stimulation of an electrode placed at that square, the positive response rate was 25%. A higher rate means that more patients produced one type of response at the stimulation of that location. Squares with positive response rates less than 10% or at least 10% of each visual response were marked differently (see figures 1 and 2⇑).
On the other hand, to find the anatomic locations frequently showing one specific visual response, the frequency of each visual response at different anatomic regions was represented by the percentage of the number of electrodes showing one type of visual response at one anatomic region to the total number of electrodes in all brain regions that produced the same type of visual response. The percentage was calculated by dividing the number of electrodes with one type of visual response at one anatomic region by the total number of electrodes with the same visual response in all brain regions, and multiplying by 100% (table).
Frequency of each visual response at different anatomic regions, represented by the percentage of the number of electrodes producing one type of visual response at one anatomic region to the total number of electrodes showing the same visual response in all brain regions
Results.
Of the total of 1,196 subdural electrodes from individual patients, visual responses were obtained by stimulation of 271 electrodes (22.7%). The mean threshold of these responses was 4.75 ± 2.33 mA (range, 2.0 to 8.5 mA). Electrical cortical stimulation of the human visual cortex produced various visual responses.
The simple forms were very small spots or blobs of flashing light that were white or black and still or moving. None of the simple forms were colored. Of the 271 electrodes with visual responses, 118 (43.5%) produced simple forms. The intermediate forms were geometric shapes such as triangles, diamonds, or stars. They were larger than simple forms, black-and-white or colored, and still or moving. The smaller forms were usually seen near the central fixation point; the larger ones were observed in the more peripheral fields. The intermediate forms occurred in 94 electrodes (34.7%). The complex forms were formed visual hallucinations such as animals, people, landscapes, and scenes from one’s memory. These complex forms were black-and-white or colored and still or moving. The stimulation of two electrodes located at the superior temporal cortex produced complex forms with auditory response (simple sound). These regions appeared to be related to multimodality perception. The sizes of complex forms were usually larger than those of simple or intermediate forms. Complex forms occurred in 59 electrodes (21.8%).
Some of the intermediate and complex forms were colored (one color or many colors, such as a rainbow). Color visual responses, regardless of shape or movement, appeared in 25 (9.2%) of 271 electrodes only. A very small, still, black spot produced after stimulation was regarded as a simple form only, whereas the electrode producing a moving diamond shape colored red was included in three types of responses—intermediate form, color, and moving sensation—at the same time.
The simple, intermediate, and complex forms sometimes moved. All visual responses with moving were classified as having moving sensation as well. Moving sensation occurred in 67 (24.7%) of 271 electrodes.
All visual responses occurred during or immediately after the electrical stimulation; they usually disappeared after cessation of stimulation but rarely persisted for a longer period after stimulation with high-intensity current, sometimes as long as 30 seconds. The most frequent visual response was a very small spot of white light. Some electrodes produced two or more small spots.
Simple form responses were seen with stimulation of the occipital pole and the striate cortex; intermediate forms with stimulation of the peristriate cortex and the lateral occipital and the fusiform gyri; complex forms with stimulation of the basal temporo-occipital and the lateral temporal or temporo-occipital junctional regions; color responses with stimulation of the basal occipital area; and moving sensations with stimulation of the basal temporo-occipital, the mesial parieto-occipital, or the temporo-parieto-occipital junctional region. The electrode locations with positive response rates less than 10% or at least 10% of specific visual response were separately mapped (see figures 1 and 2⇑). The positive response rate for a specific visual response at each square on the composite brain map was calculated by dividing the number of patients showing one type of visual response at stimulation of one square location by the total number of patients having an electrode placed and stimulated at that square, and multiplying by 100%. The stimulation of seven electrodes located on the lateral temporo-occipital or temporo-parieto-occipital junctional region showed scotoma or a large visual-field defect on the contralateral side from the stimulated hemisphere (figure 3A). Stimulation of four electrodes at the basal temporo-occipital cortex produced micropsia (shrunken image), macropsia (enlarged image), and metamorphopsia (distortion of image) in a whole visual field or a contralateral half field (figure 3B).
The distribution and frequency of subdural electrodes showing different visual responses were quantified by calculating the percentage of the number of electrodes showing one specific type of visual response at the corresponding anatomic region to the total number of electrodes in all brain regions that produced the same response (see the table). The anatomic locations frequently showing one type of visual response were listed for each response type. Simple forms were produced by stimulation of a total of 118 electrodes (56 [47.4%] were located in the occipital pole and the inferior occipital gyrus, 50 [42.4%] in the striate cortex). Intermediate forms appeared with stimulation of a total of 94 electrodes (49 electrodes [52.5%] in the cuneus and the lingual gyrus outside the striate cortex, the so-called peristriate cortex; 26 [28.0%] in the lateral occipital gyrus; and 18 [19.5%] in the fusiform gyrus). Complex forms occurred with stimulation of a total of 59 electrodes (34 [57.6%] in the basal temporo-occipital area, mainly the fusiform gyrus; and 25 [42.4%] in the lateral temporal cortex or temporo-occipital junctional region). Color responses were produced by stimulation of a total of 25 electrodes (10 [40%] in the basal temporo-occipital area, 9 [36%] in the lingual gyrus, and 6 [24%] in the inferior occipital gyrus). Moving responses were evoked by stimulation of a total of 67 electrodes (19 [28.4%] in the basal temporo-occipital area, 16 [23.9%] in the mesial parieto-occipital or temporo-parieto-occipital junctional region, 10 [14.9%] in the occipital pole; and 7 [10.4%] in the lateral posterior temporal cortex).
Discussion.
Visual processing involves parallel pathways from the retina, via the lateral geniculate nucleus and the striate cortex, finally to the extrastriate cortex. Vision seems to be mediated by at least three parallel pathways that process information for form, color, motion, and so forth. The cells in each of these visual pathways show different selectivity. In view of recent research, it seems likely that the selectivity of particular classes of neurons is related to specific aspects of visual perception. Thus, orientation-selective neurons seem to provide information for the perception of shape and form, whereas disparity-selective neurons seem to provide information about the solidity of objects. Both types of cells could be important for perceiving “what.” Direction-selective neurons concerned with motion may tell us “where.”1
In previous studies, electrical stimulation of both the mesial and the lateral occipital cortices produced simple visual hallucinations.15 These sensations were mainly single spots of white or black light at constant positions in the visual field, although for some electrodes there were two or more such spots. The size of the spot ranged from a tiny punctate sensation to the size of a large coin. These light sensations were later called phosphenes.16 Occipital-lobe stimulation most often produced phosphenes in the center of vision. Stimulation of the medial occipital region above the calcarine fissure induced phosphenes in the lower visual field contralateral to the stimulated hemisphere and stimulation below the calcarine fissure produced phosphenes in the contralateral upper visual field.11,16-18 Quadrant phosphenes were also observed if the lateral occipital cortex was stimulated.11 Other aspects of human vision have not been studied systematically.
In our study, simple forms were observed mainly if stimulation was provided to the occipital pole and the striate cortex just around the calcarine fissure. These findings imply that such areas just above and below the calcarine fissure are functionally related to the primary visual cortex. The intermediate forms were commonly produced after stimulation of the peristriate cortex, which may be related to the secondary visual cortex. Complex forms appeared if the basal temporo-occipital cortex was stimulated, usually in the fusiform gyrus and the lateral temporal or temporo-occipital cortex, which may be related to the association visual cortex for higher-order visual processing. Visual illusions such as micropsia, macropsia, and metamorphopsia were produced at the basal temporo-occipital cortex, indicating that visual distortions can be caused by stimulation of specific visual cortical areas in normal human brain. Scotoma or large visual-field defects appeared with stimulation of the lateral temporo-occipital and temporoparieto-occipital junctional regions. It is unclear whether the visual illusions and field defects result from normal visual processing or are phenomena secondary to the interruption of normal physiologic process. There are six retinotopic maps in the occipital lobe. V1 is primary visual cortex and V2 to V6 are striped regions as relay points for visual pathways. In the macaque, V2 surrounds V1. All modalities of visual stimulation activate both V1 and V2. V1 is connected with V5, both directly and through V2. V1 is also connected with V4, both directly and through V2, but mainly through V2. Functionally, monkey V4 is closely related to color response.19 Macaque V5 is activated by visual-motion stimulation.20 The topographic separation between V5 and V4 in human visual cortex is also impressive. Color responses were elicited upon stimulation of the electrodes over the basal occipital area, mainly the fusiform and the lingual gyri, which are considered the area corresponding to macaque V4. Moving sensations usually appeared if the basal temporo-occipital and the mesial parieto-occipital or temporo-parieto-occipital junctional regions were stimulated. These areas seem to be human V5. Sometimes moving simple forms appeared during stimulation of the occipital pole. In the macaque, visual motion stimuli also activate additional satellite areas surrounding V5, which are not present in human brain.20
A detailed analysis of cortical stimulation data in the human visual cortex is reported here. These data were coupled with the results from experimental studies in primates. This allows us to demonstrate direct evidence for functional organization in human visual cortex and encourages us to study more deeply how to construct the visual image in the human brain.
Footnotes
-
See also page 785
References
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
