Missing the egocentric spatial reference: a blank on the map

Introduction

Successful navigation is a fundamental cognitive function that is crucial for survival. Humans, like other animals, must learn about the layout of their environments to return home, or move between known locations. The spatial reference frame used to locate positions and directions in a complex environment are commonly divided into two main systems: egocentric (subject-centered) and allocentric (object-centered)^1,2. During navigation, information in both egocentric and allocentric reference frames can be integrated to provide a coherent representation of the environment and proper orientation.

Egocentric frames define spatial positions using the body, or a specific part of the body (head or trunk) as a point of reference (Figure 1)^3–5. Allocentric reference frame codes the position of the target relative to surrounding visual cues or landmarks and their spatial relationships, independently of the observer's current position⁶. Such information can be used to build a “cognitive map”, a sort of internal representation of the environment⁷.

Figure 1. To reach the goal (bridge on the river), the subject can reproduce a sequence of left-right body turns (egocentric strategy) or use environmental cues (allocentric strategy).

Typically, the egocentric strategy relies upon kinesthetic and vestibular sensory information as well as motor command efferent copies, and optokinetic flow information as the subject moves past surrounding objects⁸. Egocentric learning ability has been demonstrated in paradigms in which animals must repeat a sequence of responses or movements toward a target, e.g., turning to the left, or reaching a fixed goal from a fixed starting point. Egocentric spatial orientation is likely to occur in the absence of external allothetic visual cues—e.g., in total darkness or in water maze with high and opaque walls that do not allow the use of extramaze cues. However, learning occurs relatively slowly in the water maze in darkness^9,10, due to cumulative errors of the vestibular system, which are not corrected by visual inputs¹¹. Egocentric navigation is also associated with path integration, a strategy of spatial navigation that uses vestibular and proprioceptive cues generated during locomotion to keep track of position relative to a known starting point¹². Path integration or dead reckoning is, for example, used by foraging animals, such as desert ants and honeybees, to search for food along novel routes extending hundreds of meters¹³. After reaching the site, those animals show an impressive level of accuracy at returning back to the nest using only the idiothetic cues generated by their movements. Environmental allothetic cues can be used to correct heading direction. Path integration can also be assessed in human subjects walking blindfolded to a previously seen target or asking them to estimate the distance and direction traveled while walking blindfolded¹⁴.

Neuronal basis of egocentric navigation

Behavioral and brain imaging studies indicate that egocentric and allocentric strategies are mediated by a different cognitive-neural system, with some degree of overlapping. Egocentric-based navigation relies mainly on the posterior parietal cortex (PPC; Figure 2) whereas hippocampus and parahippocampal cortex are crucial for allocentric navigation^15,16. The activity in the caudated nucleus has also been associated with egocentric tasks requiring a response strategy, such as following a well-learned route in a virtual city¹⁷.

Figure 2. Brain regions involved in egocentric spatial representation.

Egocentric navigation in childhood and aging

Studies on spatial abilities in childhood suggest that infants use mainly an egocentric reference system and that a gradual ability to use allocentric representations is acquired with age. Egocentric representation is considered a more elementary mean of representing the location of an object than an allocentric representation and, therefore, is already present early^60,61. Acredolo and Evans⁶² found that 6-month-old infants, trained to anticipate the appearance of a face at either their left or right and then turned around, continued to look in the same egocentric direction after they were rotated to the opposite side of the room. The correct use of allocentric representations develops progressively with increasing age. A further study carried out in children of 5, 7, and 10-year-olds has demonstrated that a high percentage of children spontaneously used the egocentric strategy on the virtual reality adaptation of the StarMaze task, reproducing the same sequence of body turns during the probe trials as during the training trials⁶³. The allocentric strategy, based on landmark guidance, was spontaneously used to solve the task in a few percentage of children at 7 and 10 years, but not at 5 years of age. However, when the allocentric strategy was imposed, the 5- year-olds were able to use allocentric behavior but their performance was below that of the 10-year-olds.

The shift from egocentric to allocentric strategy preference with increasing age is consistent with the heterogeneity in developmental trajectories of subcortical and cortical structures; decrease of subcortical and increase of cortical gray matter volumes have been described after age 7⁶⁴.

After 60 years of age, there is a clear decline in spatial navigation abilities⁶⁵. Such decline is often related to functional changes of the posteromedial, the medial-temporal and the frontal areas^66,67. Older adults present spatial navigation deficits particularly in allocentric navigation, being less effective in forming and using the cognitive maps when examined in both virtual and real-life versions of the human Morris maze^68,69. On the other hand, the egocentric spatial navigation and learning is preserved in older age. The same results have been confirmed in studies testing older adults (71–84 years old) in a real-space human analog of the Morris Water Maze⁷⁰.

Behavioral studies on rodents have confirmed the effect of aging on the spatial task performance requiring the use of allocentric strategy⁷¹. For example, aged mice were impaired in the water maze task in which they must remember the location of a submerged platform in relation to a series of extra-maze or distal cues, with the position of the starting point changing from trial to trial⁷². However, the aged mice performed correctly the egocentric spatial task in a T-maze, remembering a sequence of movements (such as rotation to the left). These results suggest that aging can affect the allocentric and egocentric processing of spatial information in a different way.

These findings are in line with an increased sensitivity of brain regions of the medial temporal lobe, including the hippocampus, to the insult of aging and associated neurodegenerative disorders such as AD.

Egocentric spatial deficits in developmental disorders: the case of autism

Based on neuropsychological literature showing that the use egocentric develops before allocentric spatial strategies⁷³, altered egocentric spatial ability can be considered an early sign of neuro- and psychiatric developmental disorders. This issue is generally little explored. However, some studies have addressed egocentric spatial information processing in Autism spectrum disorder (ASD). ASD is a multifactorial neurodevelopmental disorder that can also be secondary to genetic syndromes, including DiGeorge syndrome, Optiz syndrome or Fragile X syndrome⁷⁴. ASD manifests with impaired eye contact and communication skills, as well as impaired social interaction⁷⁵. Furthermore, children with autism often present stereotyped behavior and self-injury. Visuo-spatial abilities in subjects with autism have been investigated. Although in one study, no impairment was observed⁷⁶, more recent studies suggested that autistic adults show impaired performance in egocentric spatial tasks, when they have to use their body as reference frame^77,78.

These deficits have never been directly associated to altered brain activation patterns. Interestingly, structural alteration of subcortical regions has been widely identified in autistic children. For example, the largest brain morphometric study in ASD to date shows that autism is associated with smaller subcortical volumes of the pallidum and putamen. Cerebellar cell loss and atrophy has also revealed in ASD children^79–83.

Impaired subcortical functions in animal models of autism have been mainly associated to social behavior impairment, with little attention to the possible effects of spatial information processing. Behavioral investigation of MID1-null mouse model of Opitz G/BBB syndrome (OS), a genetic disorder characterized by mental retardation and brain abnormalities such as hypoplasia of the anterior cerebellar vermis, has demonstrated that these mice can promptly learn to locate the food in a T-maze if the position of the food remains constant relatively to extra-maze cues (allocentric strategy), but they are impaired if the position of the food is anchored to the animal position in the maze (egocentric strategy)⁸⁴. Impairment of egocentric spatial information processing has been also observed in children with Williams syndrome (WS), a neurodevelopmental disorder resulting from a hemizygous microdeletion of ~25 genes on chromosome 7q11.23^73,85.

Egocentric spatial abilities impairment correlates with altered social performance in ASD. Therefore, understanding the nature and the neuronal correlates of egocentric spatial abilities in ASD might be relevant to shed light on the nature of other cognitive and social deficits characterizing ASD, as evidenced by studies linking spatial and social cognitive abilities in ASD^86,87.

Egocentric impairments deficits in Alzheimer's disease

Alzheimer's disease (AD) is a neurodegenerative disorder characterized by the accumulation of amyloid plaques and neurofibrillary tangle accompanied by neuronal loss⁸⁸. In the early phase of the disease, most patients show cognitive impairments, such as memory deficits, language impairment, poor judgment and decision making^89–91. As the disease progresses, reasoning, visual-spatial skills, and sensory processing become increasingly affected. The cognitive impairment in patients with AD is closely associated with the progressive degeneration of medial temporal lobe, including hippocampal formation and entorhinal cortex, frontal and parietal cortex and the basal forebrain^92,93. Spatial navigation deficits, such as becoming disoriented or feeling lost in familiar places, represent the first sign of AD, and have, therefore, an important clinical utility for the early detection of AD^94,95. Several studies in recent years have focused on spatial deficits in patients with amnestic Mild Cognitive Impairment (aMCI), which is considered an intermediate stage between the expected cognitive deficit of normal aging and the serious decline of dementia. Patients with amnestic MCI, especially those with the hippocampal type of amnestic syndrome, are at very high risk of AD. The conversion rate from MCI to dementia range is about 10% per year; in contrast to conversion rate from healthy elderly subjects to dementia which is about 1–2 percent per year⁹⁶. Interestingly, MCI patients are impaired in both allocentric and egocentric navigation, as documented by a series of study on virtual as well as real navigation. In one of the first studies, Hort and collaborators⁹⁷ tested a group of AD and aMCI patients on a goal-directed navigation task within a circular arena. The goal was invisible and could be identified either by its position relative to the starting position (egocentric subtest) or relative to cues on the wall (allocentric subtest). Each subtest began with an overhead view of the arena on a computer monitor. Both AD and aMCI were similarly impaired in all types of experiments, suggesting that spatial navigation impairment is not limited to AD, but begins to decline earlier in MCI. To understand the neural mechanisms behind spatial navigation of AD, Weniger and collaborators⁹⁸ examined a group of patients with aMCI on two virtual reality tasks, a virtual park containing several landmarks (allocentric memory) and a landmark-free virtual maze (egocentric memory) while undergoing an fMRI scan. They showed that aMCI patients were significantly impaired on both virtual reality tests, getting more frequently lost than controls and being unable to find a navigation strategy to learn the maze. Importantly, the allocentric and egocentric memory of aMCI patients were associated with size reduction of hippocampus, precuneus, and right-sided parietal cortex. These results have been recently confirmed by Boccia et al.⁹⁹. In their studies, patients with aMCI were challenged on an intensive learning paradigm, during which participants were shown two paths using either video clips or maps of a real city. The video clips were used to encourage participants to develop an egocentric representation of the city, whereas the maps were used to encourage them to develop an allocentric representation. After learning, they were asked to retrieve each of these paths, using an allocentric or egocentric frame of reference, while undergoing fMRI scan. Behavioral results indicated that aMCI showed a reduction in the rate of learning on the allocentric task. In addition, patients with aMCI showed a selective impairment in retrieving topographical memory using an egocentric perspective. Imaging data suggest that this general decline was correlated with hypoactivation of the brain areas generally involved in spatial navigation, e.g. medial temporal lobe structures, angular gyrus, and orbitofrontal cortex.

Furthermore, there is a strong evidence that patients with mild AD and aMCI are also impaired on path integration task, in which they are required to return to the starting point following an enclosed triangle pathway with a mask over their eyes¹⁰⁰. Such deficit appeared to be correlated to a reduced size of hippocampus, entorhinal and parietal cortices.

These findings suggest that neuropsychological screening designed to assess navigational deficit may represent an important tool for detecting the prodromal symptoms of aMCI and early stage of AD, and therefore, allow appropriate diagnosis and intervention.

Conclusions

Many interesting points emerge from spatial navigation studies presented in this review. First of all, it appears that, unlike allocentric, the egocentric spatial strategy is quite preserved under physiological aging. But there are conditions such as AD in which egocentric ability is also impaired and this determines devastating effects on spatial navigation so that individual get lost even in a familiar environment. Secondly, the egocentric spatial deficits in brain disorders such as AD are often correlated with a reduced size of brain areas involved in egocentric information processing. Overall, these findings suggest that neuropsychological screening designed to assess navigational deficit may represent an important approach for detecting neurological and psychiatric disease progression, thus allowing an appropriate intervention.