Scene Configuration and Object Reliability Affect the Use of Allocentric Information for Memory-Guided Reaching

Neural integration of egocentric and allocentric visual cues in the gaze system

A fundamental question in neuroscience is how the brain integrates egocentric (body-centered) and allocentric (landmark-centered) visual cues, but for many years this question was ignored in sensorimotor studies. This changed in recent behavioral experiments, but the underlying physiology of ego/allocentric integration remained largely unstudied. The specific goal of this review is to explain how prefrontal neurons integrate eye-centered and landmark-centered visual codes for optimal gaze behavior. First, we briefly review the whole brain/behavioral mechanisms for ego/allocentric integration in the human and summarize egocentric coding mechanisms in the primate gaze system. We then focus in more depth on cellular mechanisms for ego/allocentric coding in the frontal and supplementary eye fields. We first explain how prefrontal visual responses integrate eye-centered target and landmark codes to produce a transformation toward landmark-centered coordinates. Next, we describe what happens when a landmark shifts during the delay between seeing and acquiring a remembered target, initially resulting in independently coexisting ego/allocentric memory codes. We then describe how these codes are reintegrated in the motor burst for the gaze shift. Deep network simulations suggest that these properties emerge spontaneously for optimal gaze behavior. Finally, we synthesize these observations and relate them to normal brain function through a simplified conceptual model. Together, these results show that integration of visuospatial features continues well beyond visual cortex and suggest a general cellular mechanism for goal-directed visual behavior.

Scene semantics affects allocentric spatial coding for action in naturalistic (virtual) environments

Interacting with objects in our environment requires determining their locations, often with respect to surrounding objects (i.e., allocentrically). According to the scene grammar framework, these usually small, local objects are movable within a scene and represent the lowest level of a scene's hierarchy. How do higher hierarchical levels of scene grammar influence allocentric coding for memory-guided actions? Here, we focused on the effect of large, immovable objects (anchors) on the encoding of local object positions. In a virtual reality study, participants (n = 30) viewed one of four possible scenes (two kitchens or two bathrooms), with two anchors connected by a shelf, onto which were presented three local objects (congruent with one anchor) (Encoding). The scene was re-presented (Test) with 1) local objects missing and 2) one of the anchors shifted (Shift) or not (No shift). Participants, then, saw a floating local object (target), which they grabbed and placed back on the shelf in its remembered position (Response). Eye-tracking data revealed that both local objects and anchors were fixated, with preference for local objects. Additionally, anchors guided allocentric coding of local objects, despite being task-irrelevant. Overall, anchors implicitly influence spatial coding of local object locations for memory-guided actions within naturalistic (virtual) environments.

Instruction alters the influence of allocentric landmarks in a reach task

Allocentric landmarks have an implicit influence on aiming movements, but it is not clear how an explicit instruction (to aim relative to a landmark) influences reach accuracy and precision. Here, 12 participants performed a task with two instruction conditions (egocentric vs. allocentric) but with similar sensory and motor conditions. Participants fixated gaze near the center of a display aligned with their right shoulder while a target stimulus briefly appeared alongside a visual landmark in one visual field. After a brief mask/memory delay the landmark then reappeared at a different location (same or opposite visual field), creating an ego/allocentric conflict. In the egocentric condition, participants were instructed to ignore the landmark and point toward the remembered location of the target. In the allocentric condition, participants were instructed to remember the initial target location relative to the landmark and then reach relative to the shifted landmark (same or opposite visual field). To equalize motor execution between tasks, participants were instructed to anti-point (point to the visual field opposite to the remembered target) on 50% of the egocentric trials. Participants were more accurate and precise and quicker to react in the allocentric condition, especially when pointing to the opposite field. We also observed a visual field effect, where performance was worse overall in the right visual field. These results suggest that, when egocentric and allocentric cues conflict, explicit use of the visual landmark provides better reach performance than reliance on noisy egocentric signals. Such instructions might aid rehabilitation when the egocentric system is compromised by disease or injury.

Integration of landmark and saccade target signals in macaque frontal cortex visual responses

Visual landmarks influence spatial cognition and behavior, but their influence on visual codes for action is poorly understood. Here, we test landmark influence on the visual response to saccade targets recorded from 312 frontal and 256 supplementary eye field neurons in rhesus macaques. Visual response fields are characterized by recording neural responses to various target-landmark combinations, and then we test against several candidate spatial models. Overall, frontal/supplementary eye fields response fields preferentially code either saccade targets (40%/40%) or landmarks (30%/4.5%) in gaze fixation-centered coordinates, but most cells show multiplexed target-landmark coding within intermediate reference frames (between fixation-centered and landmark-centered). Further, these coding schemes interact: neurons with near-equal target and landmark coding show the biggest shift from fixation-centered toward landmark-centered target coding. These data show that landmark information is preserved and influences target coding in prefrontal visual responses, likely to stabilize movement goals in the presence of noisy egocentric signals.

Stories in the Mind? The Role of Story-Based Categorizations in Motion Classification

Categorization is fundamental for spatial and motion representation in both the domain of artificial intelligence and human cognition. In this paper, we investigated whether motion categorizations designed in artificial intelligence can inform human cognition. More concretely, we investigated if such categorizations (also known as qualitative representations) can inform the psychological understanding of human perception and memory of motion scenes. To this end, we took two motion categorizations in artificial intelligence, Motion-RCC and Motion-OPRA1 , and conducted four experiments on human perception and memory. Participants viewed simple motion scenes and judged the similarity of transformed scenes with this reference scene. Those transformed scenes differed in none, one, or both Motion-RCC and Motion-OPRA1 categories. Importantly, we applied an equal absolute metric change to those transformed scenes, so that differences in the similarity judgments should be due only to differing categories. In Experiments 1a and 1b, where the reference stimulus and transformed stimuli were visible at the same time (perception), both Motion-OPRA1 and Motion-RCC influenced the similarity judgments, with a stronger influence of Motion-OPRA1 . In Experiments 2a and 2b, where participants first memorized the reference stimulus and viewed the transformed stimuli after a short blank (memory), only Motion-OPRA1 had marked influences on the similarity judgments. Our findings demonstrate a link between human cognition and these motion categorizations developed in artificial intelligence. We argue for a continued and close multidisciplinary approach to investigating the representation of motion scenes.

Further Evidence That People Rely on Egocentric Information to Guide a Cursor to a Visible Target

Everyday movements are guided by objects’ positions relative to other items in the scene (allocentric information) as well as by objects’ positions relative to oneself (egocentric information). Allocentric information can guide movements to the remembered positions of hidden objects, but is it also used when the object remains visible? To stimulate the use of allocentric information, the position of the participant’s finger controlled the velocity of a cursor that they used to intercept moving targets, so there was no one-to-one mapping between egocentric positions of the hand and cursor. We evaluated whether participants relied on allocentric information by shifting all task-relevant items simultaneously leaving their allocentric relationships unchanged. If participants rely on allocentric information they should not respond to this perturbation. However, they did. They responded in accordance with their responses to each item shifting independently, supporting the idea that fast guidance of ongoing movements primarily relies on egocentric information.

Can ongoing movements be guided by allocentric visual information when the target is visible?

People use both egocentric (object-to-self) and allocentric (object-to-object) spatial information to interact with the world. Evidence for allocentric information guiding ongoing actions stems from studies in which people reached to where targets had previously been seen while other objects were moved. Since egocentric position judgments might fade or change when the target is removed, we sought for conditions in which people might benefit from relying on allocentric information when the target remains visible. We used a task that required participants to intercept targets that moved across a screen using a cursor that represented their finger but that moved by a different amount in a different plane. During each attempt, we perturbed the target, cursor, or background individually or all three simultaneously such that their relative positions did not change and there was no need to adjust the ongoing movement. An obvious way to avoid responding to such simultaneous perturbations is by relying on allocentric information. Relying on egocentric information would give a response that resembles the combined responses to the three isolated perturbations. The hand responded in accordance with the responses to the isolated perturbations despite the differences between how the finger and cursor moved. This response remained when the simultaneous perturbation was repeated many times, suggesting that participants hardly relied upon allocentric spatial information to control their ongoing visually guided actions.

Spatiotemporal Coding in the Macaque Supplementary Eye Fields: Landmark Influence in the Target-to-Gaze Transformation

Eye-centered (egocentric) and landmark-centered (allocentric) visual signals influence spatial cognition, navigation, and goal-directed action, but the neural mechanisms that integrate these signals for motor control are poorly understood. A likely candidate for egocentric/allocentric integration in the gaze control system is the supplementary eye fields (SEF), a mediofrontal structure with high-level “executive” functions, spatially tuned visual/motor response fields, and reciprocal projections with the frontal eye fields (FEF). To test this hypothesis, we trained two head-unrestrained monkeys (Macaca mulatta) to saccade toward a remembered visual target in the presence of a visual landmark that shifted during the delay, causing gaze end points to shift partially in the same direction. A total of 256 SEF neurons were recorded, including 68 with spatially tuned response fields. Model fits to the latter established that, like the FEF and superior colliculus (SC), spatially tuned SEF responses primarily showed an egocentric (eye-centered) target-to-gaze position transformation. However, the landmark shift influenced this default egocentric transformation: during the delay, motor neurons (with no visual response) showed a transient but unintegrated shift (i.e., not correlated with the target-to-gaze transformation), whereas during the saccade-related burst visuomotor (VM) neurons showed an integrated shift (i.e., correlated with the target-to-gaze transformation). This differed from our simultaneous FEF recordings (Bharmauria et al., 2020), which showed a transient shift in VM neurons, followed by an integrated response in all motor responses. Based on these findings and past literature, we propose that prefrontal cortex incorporates landmark-centered information into a distributed, eye-centered target-to-gaze transformation through a reciprocal prefrontal circuit.

The role of perception and action on the use of allocentric information in a large-scale virtual environment

In everyday life, our brain constantly builds spatial representations of the objects surrounding us. Many studies have investigated the nature of these spatial representations. It is well established that we use allocentric information in real-time and memory-guided movements. Most studies relied on small-scale and static experiments, leaving it unclear whether similar paradigms yield the same results on a larger scale using dynamic objects. We created a virtual reality task that required participants to encode the landing position of a virtual ball thrown by an avatar. Encoding differed in the nature of the task in that it was either purely perceptual (“view where the ball landed while standing still”—Experiment 1) or involved an action (“intercept the ball with the foot just before it lands”—Experiment 2). After encoding, participants were asked to place a real ball at the remembered landing position in the virtual scene. In some trials, we subtly shifted either the thrower or the midfield line on a soccer field to manipulate allocentric coding of the ball’s landing position. In both experiments, we were able to replicate classic findings from small-scale experiments and to generalize these results to different encoding tasks (perception vs. action) and response modes (reaching vs. walking-and-placing). Moreover, we found that participants preferably encoded the ball relative to the thrower when they had to intercept the ball, suggesting that the use of allocentric information is determined by the encoding task by enhancing task-relevant allocentric information. Our findings indicate that results previously obtained from memory-guided reaching are not restricted to small-scale movements, but generalize to whole-body movements in large-scale dynamic scenes.

Integration of Eye-Centered and Landmark-Centered Codes in Frontal Eye Field Gaze Responses

The visual system is thought to separate egocentric and allocentric representations, but behavioral experiments show that these codes are optimally integrated to influence goal-directed movements. To test if frontal cortex participates in this integration, we recorded primate frontal eye field activity during a cue-conflict memory delay saccade task. To dissociate egocentric and allocentric coordinates, we surreptitiously shifted a visual landmark during the delay period, causing saccades to deviate by 37% in the same direction. To assess the cellular mechanisms, we fit neural response fields against an egocentric (eye-centered target-to-gaze) continuum, and an allocentric shift (eye-to-landmark-centered) continuum. Initial visual responses best-fit target position. Motor responses (after the landmark shift) predicted future gaze position but embedded within the motor code was a 29% shift toward allocentric coordinates. This shift appeared transiently in memory-related visuomotor activity, and then reappeared in motor activity before saccades. Notably, fits along the egocentric and allocentric shift continua were initially independent, but became correlated across neurons just before the motor burst. Overall, these results implicate frontal cortex in the integration of egocentric and allocentric visual information for goal-directed action, and demonstrate the cell-specific, temporal progression of signal multiplexing for this process in the gaze system.

Spatial coding for memory-guided reaching in visual and pictorial spaces

An essential difference between pictorial space displayed as paintings, photographs, or computer screens, and the visual space experienced in the real world is that the observer has a defined location, and thus valid information about distance and direction of objects, in the latter but not in the former. Thus egocentric information should be more reliable in visual space, whereas allocentric information should be more reliable in pictorial space. The majority of studies relied on pictorial representations (images on a computer screen), leaving it unclear whether the same coding mechanisms apply in visual space. Using a memory-guided reaching task in virtual reality, we investigated allocentric coding in both visual space (on a table in virtual reality) and pictorial space (on a monitor that is on the table in virtual reality). Our results suggest that the brain uses allocentric information to represent objects in both pictorial and visual space. Contrary to our hypothesis, the influence of allocentric cues was stronger in visual space than in pictorial space, also after controlling for retinal stimulus size, confounding allocentric cues, and differences in presentation depth. We discuss possible reasons for stronger allocentric coding in visual than in pictorial space.

Spatial updating of allocentric landmark information in real-time and memory-guided reaching

The 2-streams model of vision suggests that egocentric and allocentric reference frames are utilized by the dorsal and the ventral stream for real-time and memory-guided movements, respectively. Recent studies argue against such a strict functional distinction and suggest that real-time and memory-guided movements recruit the same spatial maps. In this study we focus on allocentric spatial coding and updating of targets by using landmark information in real-time and memory-guided reaching. We presented participants with a naturalistic scene which consisted of six objects on a table that served as potential reach targets. Participants were informed about the target object after scene encoding, and were prompted by a go cue to reach to its position. After target identification a brief air-puff was applied to the participant's right eye inducing an eye blink. During the blink the target object disappeared from the scene, and in half of the trials the remaining objects, that functioned as landmarks, were shifted horizontally in the same direction. We found that landmark shifts systematically influenced participants' reaching endpoints irrespective of whether the movements were controlled online based on available target information (real-time movement) or memory-guided based on remembered target information (memory-guided movement). Overall, the effect of landmark shift was stronger for memory-guided than real-time reaching. Our findings suggest that humans can encode and update reach targets in an allocentric reference frame for both real-time and memory-guided movements and show stronger allocentric coding when the movement is based on memory.

Allocentric representations for target memory and reaching in human cortex

The use of allocentric cues for movement guidance is complex because it involves the integration of visual targets and independent landmarks and the conversion of this information into egocentric commands for action. Here, we focus on the mechanisms for encoding reach targets relative to visual landmarks in humans. First, we consider the behavioral results suggesting that both of these cues influence target memory, but are then transformed-at the first opportunity-into egocentric commands for action. We then consider the cortical mechanisms for these behaviors. We discuss different allocentric versus egocentric mechanisms for coding of target directional selectivity in memory (inferior temporal gyrus versus superior occipital gyrus) and distinguish these mechanisms from parieto-frontal activation for planning egocentric direction of actual reach movements. Then, we consider where and how the former allocentric representations of remembered reach targets are converted into the latter egocentric plans. In particular, our recent neuroimaging study suggests that four areas in the parietal and frontal cortex (right precuneus, bilateral dorsal premotor cortex, and right presupplementary area) participate in this allo-to-ego conversion. Finally, we provide a functional overview describing how and why egocentric and landmark-centered representations are segregated early in the visual system, but then reintegrated in the parieto-frontal cortex for action.

Facilitation of allocentric coding by virtue of object-semantics

In the field of spatial coding it is well established that we mentally represent objects for action not only relative to ourselves, egocentrically, but also relative to other objects (landmarks), allocentrically. Several factors facilitate allocentric coding, for example, when objects are task-relevant or constitute stable and reliable spatial configurations. What is unknown, however, is how object-semantics facilitate the formation of these spatial configurations and thus allocentric coding. Here we demonstrate that (i) we can quantify the semantic similarity of objects and that (ii) semantically similar objects can serve as a cluster of landmarks that are allocentrically coded. Participants arranged a set of objects based on their semantic similarity. These arrangements were then entered into a similarity analysis. Based on the results, we created two semantic classes of objects, natural and man-made, that we used in a virtual reality experiment. Participants were asked to perform memory-guided reaching movements toward the initial position of a target object in a scene while either semantically congruent or incongruent landmarks were shifted. We found that the reaching endpoints systematically deviated in the direction of landmark shift. Importantly, this effect was stronger for shifts of semantically congruent landmarks. Our findings suggest that object-semantics facilitate allocentric coding by creating stable spatial configurations.