Elias Ruokanen

Introduction

 

Spatial neglect is a neuropsychological disorder that is characterized by a deficit in shifting attention to space, most commonly to the left side of one’s visual field. A patient may, for example, be unable to consciously perceive one side of space, one side of objects or one side of their own body. The condition is particularly interesting due to the rich variety of internal dissociations it can present, and the clues those dissociations give to the underlying architecture and hierarchy of spatial representations in the human mind. In this thesis I will examine the evidence for imaginal (also called representational) neglect and whether the processes of perception and imagination have separate neural bases.

 

 

The diversity of neglect

 

Spatial neglect usually arises from damage to temporal-parietal cortical regions (the posterior parietal region has been implicated in particular) and usually affects the contralateral visual field. The most common cause is a stroke in the right hemisphere [Stone et al. (1993)]. People with spatial neglect do not necessarily suffer from perceptual disturbances per se, and their oculomotor functions may be almost entirely intact. Instead, they lack the ability to direct attention or visual search to the neglected side and fail to be aware of the content presented therein. Thalamo-cortical visual pathways may be entirely preserved, allowing for information processing in higher brain areas. Indeed, implicit processing of information in the neglected side is not an uncommon occurrence [Berti et al., 1992. Marshall & Halligan, 1988].

 

Neglect comes in many forms. Personal neglect refers to a disturbance in one’s body image and typically presents itself as a difficulty in touching the neglected side of one’s body. Hemispatial neglect refers to the neglect of a side of one’s visual field, relative to the centerline of the trunk, head, or eyes. In object-based neglect external physical objects are neglected relative to their major axes. Oliver Sacks (1985) describes a patient with damage to the “deeper and back portions of her right cerebral hemisphere”, in whom all three forms of neglect mentioned above are present: “She has totally lost the idea of ‘left’, both with regard to her own body and the world.” There is also evidence for a distinction of near and far neglect, in which neglect is only present in certain spatial distances or references [Halligan et al. (1991), Keller et al. (2005), Van der Stoep et al. (2013)]. These are usually referred to as peripersonal (within one’s reach) and extrapersonal neglect (farther than one’s reach). Imaginal neglect refers to neglect in imagined objects or scenes, of content in the mind’s eye, which can be present in the absence of any other forms of neglect [Bisiach and Luzzatti (1978), Guariglia et al. (1994), Coslett (1997)]. Imaginal neglect is more commonly referred to as representational neglect in the literature, though in this author’s view the latter term is not ideal, as it can confuse representational space with representational functions.

 

The observation that unilateral neglect may also affect representational space was first made by Bisiach and Luzzatti in a pioneering study in 1978. They presented two patients suffering unilateral neglect following a stroke, who not only neglected the left side of their visual fields, but also the left side of mental images. The patients were asked to imagine and describe a familiar scene, the Piazza del Duomo, the city square of Milan, from a certain perspective. The patients reported less landmarks that resided on the left side of their imagined view. This bias to report less left-sided landmarks, relative to the imagined viewpoint, remained when asked to imagine the scene from the opposite side of the square. The allocentric and featural information regarding the square’s layout and landmarks (usually termed as “cognitive maps”) were thus available to the patients, as long as the landmarks resided on the right side of the imagined egocentric viewpoint. The authors suggested in a later paper that imaginal neglect could arise as a defect in the generation or maintenance of the left side of mental images, where “the actual possibility of conceptualizing the existence of a left side is in fact precluded”, and that perception and imagery share a spatial representation system [Bisiach & Luzzatti (1978)].

 

An alternative interpretation was offered by Meador et al. (1987) in their study of a patient with unilateral perceptual and imaginal neglect. In a task assessing mental imagery the patient was asked to name items of an imagined scene. When the patient’s head was turned toward the left (neglected) side, he reported more items from that side than when his head was turned to the right. The researchers suggested that the imaginal neglect could be partly explained by an impairment in the ability to direct visual search to this portion of the scene. Thus the mental image is intact, but directing attention to it is not.

 

The first case of pure imaginal neglect, neglect restricted to visual imagery, an “impressive dissociation”, was reported by Guariglia et al. in 1993. They presented a 59-year-old patient who had suffered a stroke that left large lesions in the right frontal and anterior temporal lobe and the head of the caudate and anterior nuclei. The patient showed neglect for imagined scenes, but no evidence of neglect for external stimuli. The authors noted that this case study was at odds with neuropsychological findings that point to the parieto or temporo-occipital cortex as the location of mental imagery structures [Farah et al. (1989) and Kosslyn (1987, 1991)].

 

Several other cases of imaginal neglect have been reported since. Coslett (1997) reported on two patients who demonstrated a double dissociation of perceptual and imaginal neglect. One patient performed normally on standard visual tests, but failed to report items from the left side of an imagined scene. The other demonstrated substantial neglect on visual perceptual and visual attentional tasks, but had normal visual imagery. Beschin et al. (1997) presented the case of a 67-year-old man with imaginal neglect after a ischaemic stroke that damaged his right parietal lobe. Ortigue et al. (2001) reported on an 86-year-old patient with pure imaginal neglect following a right thalamic lesion. Palermo et al. (2010) presented two patients with imaginal neglect.

 

 

Working memory and navigational systems

 

Given the diverse symptomatology of neglect it is reasonable to ask whether these disorders can all be explained by a common mechanism or cause, or whether we are dealing with different processes that merely appear to be linked to one another on the behavioural level.

 

Berti & Rizzolatti (1994) presented a “unifying theory of neglect” which attempts to reconcile these different forms of neglect into a coherent whole, in which all forms are a disorder of spatial awareness caused by damage to perceptuomotor cortical and subcortical pragmatics maps (predominantly in the parietal lobe). These pragmatic maps are tied to different physical functions, such as head movements, arm movements and so on, and code for both space and movements toward the coded space. This model easily accounts for neglect dissociations, as localized brain damage will impair specific spatial “sectors” and also allows for implicit processing of information in cases of partial damage.

 

Imaginal neglect, however, poses some challenges to this model. How can damage to pragmatic maps present itself in one domain but not the other? To what extent do imagination and perception share neural bases? Several theories have been proposed in the past 30-or-so years to explain imaginal neglect, and they can be roughly divided into two groups: one focusing on working memory and attention, and the other on spatial cognition.

 

Baddeley (1981) proposed that neglect of representational space could be explained by defective visuospatial working memory, in which the “visual-spatial scratch pad” might be damaged so that “either that part of the screen was inoperative, or that the process of scanning the screen was defective”. Bisiach (1993) preferred the first alternative, deeming the second “less parsimonious, insofar as it posits two theoretical constructs – a screen and a scanning device”.

 

However, an analogue spatial representation system shared by visual perception and imagery cannot account for the dissociation between perceptual and imaginal neglect [Coslett (1997)]. Furthermore, the traditional hypothesis that working memory acts as a gateway to long-term memory runs into trouble in the presence of pure imaginal neglect; how can perceptual input “pass through” working memory correctly when said working memory is clearly impaired? A model in which perception directly activates long-term memory, the products of which are then held and manipulated in working memory, better accounts for the dissociation between perceptual and imaginal neglect [Beschin et al. (1997, 2005), Denis et al. (2002)].

 

Coslett (1997) posited that the dissociation might be explained by disruptions of forms of spatial cognition, which each compute their own spatial representation. A defective “on-line, real time” coordinate system that corresponds in part to motor or sensorimotor control would explain poor performance on visual perceptual and attentional tasks, whereas a preserved egocentric coordinate system, “a master map of space”, would allow for intact visual imagery. Ortigue et al. (2001), in their study of a patient with pure imaginal neglect, concluded that the “preservation of topographical knowledge directly indicates that representational neglect was not a result of memory impairment but rather a result of a disorder in mental imagery processes operating in a body-centered system”

 

Guariglia & Pizzamoto (2007) hypothesized two different types of space representation, “topograhical” and “non-topographical images”, that are functionally separated by their relation to navigational processes. Topographical images refer to representations of stimuli that the “subject can navigate in” and that can be “transformed into (or correspond to) mental maps of the environment”. Non-topographical images refer to representations of manipulatable objects or arrays of objects that “can never be navigated”.

 

Piccardi et al. (2008) provided evidence that damage to the cognitive system involved in coding and storing environmental information used for navigation could be the cause of some forms of imaginal neglect. Their patient showed pure imaginal neglect, but no deficits in spatial working memory “when evaluated with a traditional clinical test.” However, the patient did show “a peculiar deficit” in navigational tasks requiring re-orientation in a novel environment. The authors suggested that their patient showed damage to a “topological image system” that is devoted to mental representations of environments, as opposed to a “non-topological” system that represents static locations in the environment, such as objects.

 

This notion of topological and non-topological systems was explored by Palermo et al. (2010) in a study of two patients with imaginal neglect. The researchers subjected the patients to a battery of visual mental imagery tests and found a dissociation between environment and object mental images. The first patient, whose performance was asymmetrical on the O’Clock Test, performed poorly on tasks involving the mental generation, inspection and manipulation of objects, but showed no such deficits on tasks involving environments. The second patient, who had difficulties with the Familiar Squares Description Test, had similar problems with environments, but no difficulties with objects.

 

Further investigation into this topic was conducted by Guariglia et al. (2013), who submitted a group of 96 chronic, right brain-damaged patients to an extensive neuropsychological evaluation that included the Familiar Square Description Test and the O’Clock Test. Their results support a dissociation between topological and non-topological images, with pure imaginal neglect for non-topological images being rare, affecting about 3% of the sample of right brain-damaged patients.

 

 

The BBB model of spatial memory and imagery

 

An intriguing theory that largely falls in line with the above studies is the BBB model of spatial memory and imagery [Burgess et al. (2001), Byrne (2007)], which postulates the existence of processes of format transformation, the impairment of which results in imagery neglect. Under the BBB model both encoding and retrieval of spatial memory requires a translation between (parietal) egocentric and (medial temporal) allocentric representations. These processes of translation are mediated by posterior parietal and retrospenial areas.

 

According to the model, visual imagery on the products of retrieved information from memory and sensory perception both occur in the egocentric frame or “parietal window” [Byrne (2007)]. The hippocampus and surrounding medial temporal lobe structures index information of real or imagined locations and Papez’s circuit translates to imagery or from perception according to the direction of view.

 

Damage to these connections could explain difficulties in the Familiar Squares Description Test, as the stored allocentric information would not be correlated with the egocentric point of view. It also accounts for imaginal neglect in cases of anterior cortex [Guariglia (1994)] or thalamic [Ortigue et al. (2001)] lesions, as the retrosplenial cortex has plenty of connection with both. It predicts the effect of manipulating a patient’s perceived head direction on the retrieval of the spatial context of episodic memories [Meador et al. (1987)]. Pure perceptual neglect in the absence of imaginal neglect could be explained by “a lesion along the sensory or motor pathways projecting into and out of posterior parietal cortex”, while pathways to and from the medial temporal lobe remain intact. Pure imaginal neglect would have the opposite neuroanatomical configuration. Furthermore, the BBB model is more in line with the unifying theory of neglect, proposing clear neuroanatomical causes and not relying on more abstract concepts of working memory as Baddeley and Lieberman (1981) and Bisiach (1993) do.

 

Palermo et al. (2012) performed a study based on the BBB model in which patients with perceptual neglect and imaginal neglect had to memorize a novel environment (a hospital ward they had not explored before) and draw a map of it. In another task they had to learn a map of a convention center and then navigate inside the real building. The patients with imaginal neglect had very poor allocentric representations of the hospital ward, as showed by their drawings, which had few details. In the second task both perceptual and imaginal neglect patients had great difficulty navigating the environment, but perceptual neglect patients only made errors when left turns were required, whereas patients with imaginal neglect failed to reach their target when the perspective inside the environment did not match the orientation of the learned map, regardless of whether left or right turns were required. These difficulties with transformation from egocentric to allocentric and vice versa fit with the BBB model of spatial memory and imagery.

 

Boccia et al. (2018) investigated whether damage to white matter tracts in the parietal lobe would prevent the integration of egocentric and allocentric information and result in imaginal neglect for places. They examined 58 right brain damaged stroke patients with and without extrapersonal perceptual neglect and/or imaginal neglect for places using MRI scans. They found that patients with imaginal neglect “for places” showed a higher probability of damage to the posterior part of the arcuate fascilicus and the first branch of the superior longitudal fascilicus. Disconnections in different parts of these areas have been implicated in extrapersonal perceptual neglect [Thiebaut de Schotten et al. (2014)]. They also found that damage to sections of fronto-parietal u-shaped tract predicts imaginal neglect for places, even when extrapersonal neglect is taken into account, but caution that interpretation of its role "is difficult and deserves some caution”.

 

 

Conclusion

 

Spatial neglect provides an illuminating glimpse at the structures of perception and imagination and how mental representations are generated, maintained and investigated. The diverse internal dissociations have presented a conundrum that has taken decades to unravel into a promisingly coherent picture.

 

The separation of spatial representations into topological and non-topological systems appears to this author as a milestone, foregoing more abstract models of spatial working memory for two concrete, functional categories that fulfil different pragmatic goals (navigation of environments and manipulation of objects). The BBB model that is built on these spatial representation systems and the connections between them appears to explain the internal dissociations of neglect and provides neuroanatomical foundations for them.

 

However, the understanding of neglect is confounded by the fact that it is largely built on lesion studies, and it is the nature of lesions to not be confined to specific areas for the convenience of scientists. Multiple areas and functions of the brain are usually damaged, which precludes one from drawing straightforward conclusions about the underlying nature of these phenomena. The combination of transcranial magnetic stimulation and virtual reality might provide a way to study this phenomena in a more precise manner with healthy individuals. Both have already been used on their own to study neglect, but have never been combined [Burgess et al. (2001), Bjoertomt et al. (2002), Ricci et al. (2012)].