• We use a slowly expanding ring to map retinotopy in the eccentric dimension. The field size, check size and check color can vary. We typically use a 25% duty cycle so that 1/4 of the screen is filled with checks at any moment in time.
• We use a slowly rotating wedge to map retinotopy in the angular dimension. We typically use a 90 deg wedge that sweeps out a field comparable to that mapped by the expanding ring.
• We typically scan for 10 expansion or rotation cycles of the stimulus, with a cycle period of 24 seconds. I.e., so that any given neuron is stimulated for ~6 secs at a time (25% duty cycle).
• We typically do half the scans expanding and half contracting. Likewise, we do half clockwise and half counter-clockwise.
• We have explored a number of different ring and wedge sizes, along with different cycle periods, but this has probably not been completely optimized.  In addition, the optimal choices may vary for different visual areas.

Example expanding ring and rotating wedge stimuli

 

• The initial step of the data analysis is to average the data across repeated wegde scans and separately across repeated ring scans. This involves time-reversing the contracting rings and counter-clockwise wedges, and time-shifting to compensate for the hemodynamic delay. This is explained in the mrLoadRet user's guide .
• You must also segment the gray matter using mrGray and flatten the cortex using mrFlatMesh .

Inplane representation:

• Run mrLoadRet in the directory containing your retinotopy data.
• Load the inplane anatomies, choose the averages view type, and view the phase map.  Click on the scan number corresponding to an expanding ring scan. Try setting the correlation threshold around 0.50.

Below is an example of phase data on two separate inplane images from a subject viewing a rotating wedge (90 deg wedge, 36 sec/cycle, 6 cycles/scan, from /usr/local/mri/retinotopy/baseler/101598). Inplane images with functional phase data (colors) superimposed ( rotating wedge ) :

• Determine dorsal and ventral:
  Select a rectangular ROI in the flattened representation on one side of the calcarine, then transform it to the inplane or volume view.  Ventral is the side closest to the cerebellum on the bottom of the brain.  (Note that the inplane view can be upside-down). Dorsal is the opposite side, closer to the top of the head.

ROI on flattened left hemisphere and inplane

• In the Flat window, view the phase map for the expanding ring data.  Adjust the correlation threshold to strikes a a balance, displaying contiguous bands of phase/color (good data) and cutting out random, fragmented (noisy data). The choice of correlation threshold will depend on the number of repeated scans that you have. Blurring the data can also be very helpful, especially for less experienced users.
• Phases for these data represent eccentricity in the visual field.  The eccentric representation generally runs from posterior to anterior in all retinotopic visual areas in the brain, from the occipital pole in the back of the head towards the front of the head, parallel to the calcarine sulcus.  This means that the iso-eccentricity contours (bands of constant color/phase) run roughly perpendicular to the calcarine sulcus .
• In a normal subject, the largest clump of contiguous color or phase should represent the
• fovea , or center of the visual field, where the subject was fixating.
• The phase colors should progress from the foveal phase in an orderly fashion from left to right on the mrLoadRet colorbar.  The most anterior contiguous band of color represents the most eccentric position, or the edge of the stimulus.  Beyond this band, the phases or colors may become fragmented and random.

• In the Flat window, view the phase map for the rotating wedge data.  I find that a correlation threshold of around 0.15-0.20 works best.  Again, blurring the data can be very helpful, especially for less experienced users.
• Phases for these data represent polar angle around the visual field. To help orient yourself and to identify regions of constant polar angle, contrast the angular map with the eccentricity map.  Note that the bands of constant phase/color, or iso-angular contours , run perpendicular to the iso-eccentricity contours.  This also means that the iso-angular contours run roughly parallel to the calcarine sulcus .
• Notice that the angular map only represents one hemifield , i.e. it only contains half of the phases/colors in the colormap.  The visual field is left-right reversed in the brain; the left hemisphere contains a representation of the right hemifield, and the right hemisphere represents the left hemifield.
• The angular representation is also inverted in the brain.  The upper visual field is represented in the lower part of the brain ventral/inferior to the calcarine sulcus, while the lower visual field is represented in the upper part of the brain dorsal/superior to the calcarine sulcus.  This means that the ventral side of the flat map should contain mostly colors confined to one quarter of the colormap, representing one quadrant of the hemifield, and the dorsal side should contain the other quarter (half of the colormap for this hemisphere).

• Continue viewing the phase map for the rotating wedge data in the Flat window.
• V1:   The upper and lower quadrants should meet at the representation of the horizontal meridian, near the calcarine sulcus.  Therefore, there should be the representation of one entire hemifield (180 deg of phase, or half the color map) centered on the calcarine sulcus.  This is primary visual cortex, or area V1 .  The boundaries of V1 are located at the representation of the vertical meridian .  The upper boundary of V1 is formed by the representation of the lower vertical meridian, and the lower boundary of V1 is formed by the representation of the upper vertical meridian.
• V2:   At the V1 boundaries, the orderly progression of colors reverses direction. On either side, the colors now move from the representation of the vertical meridian back to the representation of the horizontal meridian, forming the representation of either the upper or lower visual quadrant.  The lower quadrant representation on the dorsal side of the map is called V2d (dorsal) .  The representation of the upper quadrant on the ventral side of the map is called V2v (ventral) .
• V3:   The orderly progression of colors reverses direction again at the representation of the horizontal meridian .  The horizontal meridian representations form the boundaries between V2 and V3 .  The representation of the horizontal meridian on the dorsal side forms the boundary between V2d and V3d , and on the ventral side between V2v and V3v .
  V3a:  On the dorsal side of the brain, just beyond area V3d, there is another representation of a full hemifield , rather than just a quadrant.  This is area V3a . Although the unfold of the left hemisphere in the example below does not quite capture V3a, you can see it in the example of the flattened right hemisphere.  V3a shares its lower boundary with area V3d at the representation of the lower vertical meridian, and its upper boundary represents the upper vertical meridian.  (Note that this in violation of the general principal of the inverted visual field representation in the brain). V3a can often be found in a deep sulcus on the dorsal surface of the occipital lobe called the transverse occipital sulcus (TOS) (Tootell et al., J. Neurosci., 1997).

Flattened left hemisphere Flattened right hemisphere

• V4v:  On the ventral side of the brain, just beyond area V3v, there is another representation of at least the upper quadrant , but often of the full contralateral hemifield .  This is area V4v .  Unlike all of the other retinotopic areas mentioned so far, V4v does not have a clear homolog in the macaque brain.  The function and retinotopic representation of several  areas on the ventral surface of the human brain (V4v,V4,V8) are still being debated in the literature (McKeefry & Zeki, 1997; Hadjikhani et al., 1998; also, cf. Engel et al., 1997b).

Volume slice representation: Once visual areas have been identified, you can create polygonal ROIs on the flat maps to represent each visual area and transform the ROIs to the volume representation in mrLoadRet , or transform them to ROIs that can be read and visualized by mrGray . The figure below shows three slices from different orientations.  Each visual area is a color-coded ROI overlay displayed in mrGray.  The color code for each visual area is shown on the right.

• Below are three-dimensional renderings of the left occipital lobe and part of the attached parietal lobe from the subject whose retinotopy data are shown above.  The cortical surface was rendered using mrGray near the boundary between the white and gray matter (specifically, between layers 1 and 2 in the gray matter).  Two views are shown of the same rendering -- a posterior view and a medial view.
• Visual areas:  Visual areas, selected from the flattened representation and transformed as  ROIs to mrGray, are superimposed and color-coded according to the legend on the left. Notice that V1 falls within and around the calcarine sulcus, and that the other retinotopic areas fall on either side of V1.

• Eccentricity map:   The colored functional data below show the representation of eccentricity on the same observer's rendered left occipital lobe.  The color code for eccentricity is shown in the legend on the right.  Notice the large region devoted to the foveal representation (shown in blue) on the lateral surface!

References you should have and cite: