Cortical Plasticity in Adult Human Vision

Does visual cortex change in adulthood? And, if so, how?

Many have argued that early visual cortex, once wired up in development, remains fixed in adulthood. However, my research challenges this hotly-debated belief. My research addressing this question began with my Ph.D. work on a 50 year-old stroke patient (BL), 6 months post stroke. Patient BL has intact primary visual cortex (V1), but his stroke destroyed the optic radiation fibers that provide input to the V1 region representing the upper left visual field, leaving him blind in his upper left visual field. Critically, I found with fMRI that the region of V1 normally activated only by sitmuli in the upper left visual field became responsive to stimuli in the lower left visual field, demonstrating "reorganization" of V1 following deprivation (Dilks et al., 2007) [MIT News, Neurology Today].

Our current research is investigating the causes ( Dilks et al., 2014; Baker et al., 2008) and underlying neural mechanisms (Dilks et al., 2009) of cortical reorganization in adulthood, by testing another much larger population of adults with prolonged deprivation (years) of early visual cortex: individuals with macular degeneration.

How does reorganization of visual cortex affect perception?

The reorganization I have reported in stroke patient BL and in adults with macular degeneration raises a fundamental question for cognitive neuroscience: How does cortical change affect perception? What does patient BL "see" when a stimulus is presented in his intact lower left visual field, below his blind area? I found that BL experiences perceptual distortion as a consequence of cortical reorganization: a square presented in the lower left visual field was perceived as a rectangle extending into the blind upper visual field (Dilks et al., 2007) (see Figure below).

Our current research, using psychophysics and TMS, is asking whether similar distortions are present in individuals with macular degeneration. We are testing the prediction that perceptual distortion analogous to what I found in BL will be seen in individuals with macular degeneration who show reorganization with fMRI , and not in those who do not show such reorganization.


An example of BL's perceptual distortions as a consequence of cortical reorginization

When a stimulus appears just below BL’s blind area, the shape elongates upwards and into the blind area. His drawings here depict that he perceived circles as cigar shaped, squares as rectangles, and triangles as pencil shaped.

Not only was perception affected in BL, but also visually-guided action. Here are two short videos showing BL reaching out for shapes in his unaffected lower right visual field (showing normal behavior), and in his affected lower left visual field (showing the distortion as he reaches for a rectangle that he perceives to be larger than it is).


How quickly can deprived cortex change?

The time course of cortical change can provide crucial clues into the underlying neural mechanisms. I therefore devised a novel paradigm to chart the time course of cortical change following deprivation in normal adult humans. How quickly can the deprived cortex change? The key innovation enabling me to test this question was to patch one eye in normal observers, thus depriving the cortical region corresponding to the other eye's blind spot. I then tested for perceptual distortions by probing, at various time intervals after the onset of patching, for perceived elongation of shapes presented adjacent to the deprived location. I found that significant elongation occurred around the blind spot within seconds of patching, indicating very rapid reorganization (Dilks et al., 2009) [Scientific American]. These results implicate unmasking of pre-existing connections as the underlying neural mechanism of such rapid reorganization, rather than the growth of new connections.

Could this rapid change reflect the same underlying neural mechanism as changes in early visual cortex after months or years as shown in patient BL and individuals with macular degeneration? Ongoing research is testing the hypothesis that cortex can change on two time scales: i ) short-term (within seconds, possibly related to GABA, which we are now testing with magnetic resonance spectroscopy - MRS, and pharmacological manipulations), and ii) long-term (months or years, testable by comparing the blind spot in participants whose cortical representatino of the blind spot has been deprived for years because of the loss of an eye).


Other questions under investigation...

How does higher-level cortex change in adulthood?

Does cortical reorganization play some directly useful role in adult vision?

How does deprivation affect the developing cortex, and how is it different from cortical reorganization in adulthood?



Functional Organization of Human
Visual Cortex and its Origins

What are the functional properties of the cortical regions involved in visually perceiving people, places, and things in adulthood?

Studies of adult humans and monkeys have discovered a number of neural structures selectively responsive to complex visual stimuli, including faces, places, words, and objects (Kanwisher & Dilks, 2013; Dilks et al., 2013; Dilks et al., 2011) (see Figure below). However, for the discoveries of functionally specific brain regions to be useful in understanding the human mind (our goal as cognitive scientists), we need a much richer understanding of the role of each of these regions in cognition. We need not just loose descriptions of the function of a region (e.g., face perception) but precise characterization of the computations and representations conducted in each region. For example, we found that a face-selective region in the posterior superior temporal sulcus (pSTS) responds about three times as strongly to movies of faces (but not movies of bodies or objects) as to static snapshots taken from those face movies, whereas another face-selective region - the fusiform face area (FFA) - responds the same to movies and snapshots (Pitcher et al., 2011). These findings implicate the pSTS in the representation of more dynamic high-level face and social information, including eye, mouth and head movements and facial expression than the FFA. Similarly, for scene perception, we recently found that representations in one scene-selective region - the parahippocampal place area (PPA) - are largely invariant to mirror-image reversals of scenes, while two other scene-selective regions (the retrosplenial complex, RSC, and the occipital place area, OPA, formerly referred to as TOS; (Dilks et al., 2013) are sensitive to such mirror-image reversals (Dilks et al., 2011). This finding challenges the hypothesis that the PPA is directly involved in navigation and reorientation and suggests instead that scenes, like objects, are processed by distinct pathways guiding recognition and action.

Ongoing work is further investigating the representations and computations of each of the specialized regions of cortex, as well as how they are connected (Zhu et al., 2011).


Specialized Regions of Visual Cortex

Inflated brain from three representative subjects showing regions specifically involved in the perception of faces (blues), places (pinks), and bodies (green). Dark blue = FFA; Purplish blue = pSTS; Light blue = OFA. Magenta = PPA; Light purple/pink = RSC; Reddish pink = OPA; Green = EBA. Each of these regions can be found in a short functional scan in essentially every healthy subject. LOC is not shown here; It is a very large region generally responsive to any object shape, and hence both of its subregions (LO and pFs) overlap partially with some of the regions shown here. The VWFA (left hemisphere) and FBA (partially overlapping with FFA) are also not pictured.


How does functional specialization get wired up in development?

How does functionally specialized visual cortex get wired up? Face recognition offers an important test case: While several papers have claimed that one face-selective region - the fusiform face area (FFA) - continues to grow into the teenage years, suggesting a role for prolonged experience in the development of face processing mechanisms, the key behavioral signatures of face processing implicating the FFA are adultlike by age 4 (and many are present in infancy). Thus, a puzzle arises: If the behavioral abilities are in place by age 4, why is the brain still changing after that? We hypothesized that the apparent change in the FFA may be due to the technical challenges of scanning young children, and the weaker data available (and smaller FFAs) that result. To test this hypothesis, we took two steps to substantially improve pediatric neuroimaging methods by: i) devising novel, highly engaging experimental stimuli, and ii) collaborating with MRI physicists to design a special pediatric head coil (Keil et al., 2011). With these improved methods, we see an adult-like FFA in five to six year olds (see Figure below and a talk I recently gave on "What can brain imaging tell us about the developing child's brain?"). Demonstrating that the FFA is present as early as 5 years of age, this result reconciles the conflict between the neuroimaging and behavioral literatures, and it underscores the importance of optimizing methods in pediatric neuroimaging studies.

Although the above finding provides an upper age limit for the development of face recognition, the questions of when the machinery underlying adult-like face recognition arises remains. Current work is addressing this question by scanning yet younger children (infants), using a custom-made coil for infants. Scanning infants will not only allow us to identify when the machinery underlying adult-like face perception arises, but will also critically enable us to investigate how the neural machinery of face perception changes over the entire course of development. Ongoing work is also examining the developmental profiles of scene recognition.

An example right (circled) and left hemisphere FFA in two typical adults and two typical 5-6 year olds (who were scanned in the custom-made pediatric head coil), labeled accordingly.


What is the role of genes in orchestrating the development of the functional architecture of the human visual cortex?

Ultimately, however, we need to know not only when, but also how functional specialization arises in development. What is the role of genes in orchestrating this development? To begin addressing this question, I investigated the development of the dorsal and ventral streams in individuals with Williams syndrome (WS) - a rare genetic disorder. Specifically, I compared the performance of individuals with WS to typically developing children at various ages on two tasks: a vision-for-action (dorsal stream) task and a vision-for-perception (ventral stream) task. I found a selective impairment in the action task compared to the perception task in the WS individuals relative to mental-age-matched controls (Dilks et al., 2008; Landau et al., 2006). These findings show that some genes play cognitively and neurally specific roles in development. Further, by testing typically developing children at various ages (beginning at 3 years old), I found that the dorsal stream is slower to develop than the ventral stream (Dilks et al., 2008). These findings not only support the hypothesis of functionally distinct visual streams for action and perception, but also crucially demonstrate that the two streams follow different developmental trajectories.

Ongoing work uses behavioral and neuroimaging methods to examine genetic influences on the development of other specialized functions (i.e., face, place, and object recognition) in individuals with WS and Autism.