Visual imagery in cerebral visual dysfunction

Visual imagery in cerebral visual dysfunction
of 16
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Visual imagery in cerebral visual dysfunction Giorgio Ganis, PhD a,b, *, William L. Thompson, BA a ,Fred W. Mast, PhD c , Stephen M. Kosslyn, PhD a,d a Department of Psychology, Harvard University, 33 Kirkland Street,Cambridge, MA 02138, USA b Department of Radiology, Massachusetts General Hospital, Building 149,13th Street, Charlestown, MA 02129, USA c Institute of Psychology, University of Zurich, 8006 Zurich, Switzerland  d Department of Neurology, Massachusetts General Hospital,55 Fruit Street, Boston, MA 02114, USA Visual perception takes place when an object or event is being viewed,which leads to the construction of visual representations used in recognition,identification, and navigation. In contrast, visual imagery takes place whena short-term memory (STM) representation of an object or event is presentbut the stimulus itself is no longer being viewed. Visual mental imagescan be conceptualized as STM representations that are associated with theexperience of ‘‘seeing with the mind’s eye.’’ Unlike afterimages, the visualrepresentations underlying imagery are relatively extended in time.Considerable progress has been made since the mid 1980s in understand-ing the cognitive and neural processes supporting visual imagery. Severallines of research have provided converging evidence for two main principles:(1) visual imagery is not a unitary, undifferentiated faculty, but is the resultof the operation of many processes and (2) visual imagery and perceptionshare some processes and structures [1–3], but not all. Most of the researchin the field can be characterized as efforts towards refining these two princi-ples.Although neuroimaging studies provide considerable evidence in supportof both principles, it is not always clear how to interpret the available data.For instance, using functional magnetic resonance imaging (fMRI) to showthat the primary visual cortex is activated by visual imagery and perceptionis not sufficient to demonstrate that the primary visual cortex is necessary in Neurol Clin N Am 21 (2003) 631–646*Corresponding author. Department of Psychology, Harvard University, 33 KirklandStreet, Cambridge, MA 02138, USA. E-mail address:  ganis@wjh.harvard.edu (G. Ganis).0733-8619/03/$ - see front matter    2003 Elsevier Inc. All rights reserved.PII: S 0 7 3 3 - 8 6 1 9 ( 0 2 ) 0 0 0 9 7 - X  both cases or that it is necessary for the same reasons [4]. Evidence fromneurologic cases potentially can resolve much of this type of ambiguity[5]. For instance, if brain region X is activated by visual imagery and percep-tion but this activation does not reflect a process that actually plays a func-tional role in task performance, during imagery, then in principle damageto region X should result in impaired visual perception but not imagery.Thus, evidence from neuropsychology can be complementary to that ob-tained from the study of neurologically intact individuals [6].In practice, however, it is important to keep in mind that the interpreta-tion of neuropsychologic data is complicated by a number of factors: (1) Spatial scale:  Functions X and Y can be localized in adjacent regions oreven interdigitated at very small spatial scales. The closer the two regionsthat support functions X and Y, the more likely they will be damagedtogether by lesions, leading to spurious patterns of common impairmentof X and Y. A similar problem is present in functional neuroimaging stud-ies, where common regions of activation might reflect distinct neural pro-cesses segregated at spatial scales smaller than the size of the imagingvoxels. (2)  Reorganization and compensation:  Function X may rely on brainregion A, but X may seem not to be impaired after damage to A, becauseother brain regions can compensate for the lesion. Some forms of compen-sation probably are responsible for situations in which a function graduallyis recovered after brain damage. (3)  Localization:  Brain damage is rarelyfocal and often involves structures that are functionally unrelated. Further-more, it is difficult to assess the true extent of a lesion, because functionaldamage may not be visible on MRI scans. Also, the mapping of functionto neuroanatomic landmarks is variable across individuals. The amountof ‘‘noise’’ introduced by these factors in the data currently is not known,but most likely it is one of the reasons for the large variability in the liter-ature. (4)  Difficulty:  In general, brain damage slows down processing anddecreases accuracy, and these effects are more severe for more taxing tasks.Thus, a single dissociation in performance between two tasks after braindamage could reflect simply the fact that one task is more difficult overallthan the other. To rule out this possibility, researchers must demonstratethat some tasks of equal overall difficulty (as assessed in control subjects)remain unimpaired—or at least are less impaired—than the task of interest.They also can show that some patients exhibit the opposite pattern of deficits (ie, demonstratea double dissociation).Deficits in visual mental imagery after brain damage initially were docu-mented many years ago. The first such case was a man described by Charcotwho became unable to create visual images as a result of a stroke [7].Unfortunately, the kind of tests necessary to evaluate current cognitivemodels of visual imagery, and how they relate to models of visual percep-tion, were not conducted before the 1980s, and are still conducted only spor-adically [8]. This is not surprising because the particular tests given to apatient are determined not only by practical constraints, but also by the 632  G. Ganis et al / Neurol Clin N Am 21 (2003) 631–646   specific interests of the clinician or investigator (eg, the investigator’s implicitor explicit conceptualization of mental imagery).The most articulated model of visual imagery probably is the one pro-posed by Kosslyn (see Fig. 1) [2,9]. Because this model has not only beenused extensively to interpret patterns of neurologic deficits, but also to guidesubsequent research [3], it is described briefly (this model is discussed exten-sively in Kosslyn [2]). In its simplest form, visual imagery in this model canbe thought of as running visual perception in reverse (itself an old idea).During visual perception, external images are projected onto the retina andpreprocessed by the  visual buffer , composed of retinotopically organizedvisual areas in the occipital lobe. A subset of the information in the visualbuffer is selected by  attentional processes  and sent forward for further pro-cessing to two systems: the  object-properties  system, specialized for process-ing aspects of objects such as shape and color, and the  spatial-properties processing   system, specialized for processing characteristics such as locationand orientation. The object-properties processing system is implemented bya pathway running ventrolaterally from the occipital lobe to the inferiortemporal lobe. The spatial-properties processing system is implemented bya pathway running dorsally, from the occipital lobe to the posterior parietallobe. Information from these two systems then reaches  associative memory ,where multimodal knowledge about objects is stored. A short-term associa-tive memory structure is implemented in dorsolateral prefrontal cortex,whereas a long-term associative memory relies on classical ‘‘association cor-tex’’ in the occipital/temporal/parietal junction area. The  information shunt-ing system , implemented by prefrontal cortices, accesses knowledge stored inassociative memory during top-down hypothesis testing, which involvessearching for additional information within the image. Such a search isguided by hypotheses derived from knowledge stored in memory. For ex-ample, upon seeing a wing-like shape, one hypothesis is that the imagecorresponds to an airplane, leading one to search for another distinctiveairplane part, such as a tail or wheel. Fig. 1. Major subsystems used in high-level vision and mental imagery. ( Adapted from  KosslynSM. Image and brain. Cambridge (MA): Harvard University Press; 1994; with permission.)633 G. Ganis et al / Neurol Clin N Am 21 (2003) 631–646   Top-down hypothesis testing is achieved by shifting attention to relevantparts of the visual field and by priming the corresponding visual representa-tions in the object- and spatial-properties processing systems.According to this model, during visual imagery the information shuntingsystem accesses the stored representation of the structure of an object inlong-term associative memory and sends information to the object- andspatial-properties processing systems to activate a representation of visualproperties. This activation process is identical to the priming that occursduring top-down hypothesis testing in perception; however, now the primingis so strong that activation propagates backwards, and an image representa-tion is formed in the visual buffer. Shapes and spatial relations can beinspected and identified in an image by the same mechanisms used to inspectobjects and locations during perception. Consistent with this idea is evidencethat people can reinterpret patterns in mental images, as they do with per-ceived images, provided that capacity limitations are not exceeded [10–14].Even this brief description should make it clear that deficits in visualimagery may arise from damage to several brain regions (consistent with thefirst principle described at the outset). For instance, according to this model,deficits in generating (creating) visual mental images can result from damageto the information in long-term associative memory, to the informationshunting system in the frontal lobes, to the object-properties processingsystem, and to the visual buffer. Historically, researchers have focused oneffects of damage to the brain areas that implement the visual buffer and theobject-properties and spatial-properties processing systems. These structuresalso are the focus of this article. Effects of damage to the occipital lobe If the visual buffer is necessary for visual mental imagery, then damage tothe retinotopically organized visual areas that support it should impair atleast some aspects of visual imagery. Lesions in primary visual cortex, thefirst visual area to receive information from the lateral geniculate nucleus,impair visual perception in specific regions of space. Circumscribed unilat-eral damage to primary visual cortex gives rise to hemianopsia or scotomas,whereas extensive bilateral damage to primary visual cortex (eg, as a conse-quence of infarcts to the posterior cerebral arteries) may result in corticalblindness. The ideal way to determine definitively that a specific early visualarea is damaged is to perform visual retinotopic mapping using fMRI in thesame individual before the lesion occurs. Even in this case, it is difficult tostate with certainty that a given area was not affected by the lesion, becausenormal appearance on an MRI scan does not guarantee normal functioningtissue. Therefore, the localization of a lesion to a particular visual arearemains tentative and is based, at best, on anatomic landmarks (eg, the cal-carine fissure as a definition of primary visual cortex). Also, damaging theconnections among areas can disrupt processing as much as damaging the 634  G. Ganis et al / Neurol Clin N Am 21 (2003) 631–646   areas themselves, and good methods of precisely documenting damage towhite matter have yet to be developed.With these caveats in mind, researchers have reported visual imagery def-icits in patients with unilateral occipital infarcts that produce hemianopia.The primary study with detailed testing of visual imagery was reported byButter et al [15]. These researchers tested a group of eight hemianopicpatients on a visual-image scanning task. The patient had to judge whetheror not an arrow was pointing at one of a set of dots recently seen, but whichwere no longer visible. The patients made more errors when the arrowpointed to the side ipsilateral to their hemianopia than when the arrowpointed to the side contralateral to their hemianopia. This design wasappealing because individuals served as their own controls. Because no neu-roanatomic scanning was obtained for three of the eight patients, however,and only CT scans were obtained for the remaining five, it is impossible totell whether or not the lesion involved only primary visual cortex.In addition, Farah et al [16] examined the extent of images in a patientbefore and after resection of one occipital lobe. Consistent with the findingsof Butter et al [15], Farah et al [16] found that the horizontal ‘‘visual angle’’subtended by objects in images was reduced by approximately half after thesurgery. This finding suggests that half the horizontal extent of the visualbuffer was removed by the surgery. In contrast, the vertical extent of objectsin images remained the same. Given that vertical extent is represented inboth occipital lobes, this finding makes sense and also provides a controlagainst the possibility that the brain damage simply had made the task over-all more difficult.In contrast, some researchers have reported patients cortically blind fromextensive bilateral damage to primary visual cortex who seem to have pre-served visual imagery. In one oft-cited report, Chatterjee and Southwood[17] studied a 29-year-old woman who was cortically blind from a stroke shehad suffered 14 years earlier. The damage seen on CT scan included primaryvisual cortex bilaterally. Visual perception was disrupted severely in thispatient (she was not even able to tell if a penlight was being flashed in hereyes), yet she could answer visual imagery questions ‘‘accurately and with-out hesitation.’’ The questions used include, ‘‘Is the capital letter D formedby straight lines, curved lines, or a combination of both?’’; ‘‘The body of asnake usually appears curved—true or false?’’; and ‘‘Which two of thesethree items are similar in shape: pliers, clothes pin, and paper clip?’’ [17].This patient could draw from memory the six objects the investigatorsnamed (butterfly, car, flower, dog, face, and person). Furthermore, after theonset of her blindness, she finished her high school studies and spontane-ously observed that she got through school by visualizing her studies. Othercases of cortically blind patients with preserved imagery have been reported[18], although typically the tests of visual imagery have been rather cursory.One possible explanation for preserved imagery in cortically blind indi-viduals is that ‘‘islands’’ of intact cortex are responsible for the spared 635 G. Ganis et al / Neurol Clin N Am 21 (2003) 631–646 
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks