Spatial perception is essential for sensing and interacting with the environment. During development, coherent multisensory representations of space are created by aligning and integrating information across sensory modalities. In sighted adults, vision is typically the most reliable spatial sense and has a key role in multisensory spatial development by offering an immediate and comprehensive representation of the surrounding layout in a single glance. In this Review, we explore how sensory and motor systems develop in early life to form spatial representations of the external environment, and the role of visual experience in this process. We examine differences in spatial perceptual development between visually impaired individuals and sighted individuals, focusing on visual, auditory and tactile modalities and their interactions in the context of sensorimotor and bodily changes. Differences between visually impaired individuals and sighted individuals illustrate critical developmental periods during which sensory and motor experiences influence the construction of spatial representations. In sighted adults, spatial perception relies on information integrated across sensory modalities. In this Review, Gori and colleagues consider spatial perception in visually impaired infants, children and adults to clarify how it develops and the role of sensory information in the process.

Adults are able to orient their attention between different (often fleeting) sensory cues that present to different senses ("crossmodal attention"), but we cannot assume that this is the case in early life. Here we report the findings of a study in which we probed the presence of tactile to visual crossmodal links in spatial attention using event related potentials (ERP) gathered from the scalp electroencephalograms (EEG) of 5- (n = 19) and 8-month-old (n = 19) infants. Whilst recording EEG, we presented 5- and 8-month-old participants with vibrotactile stimuli on one of their hands, followed by visual stimuli on the same (Congruent) or the opposite (Incongruent) hand, or no touch at all (No probe). During the presentation of these stimuli the infants' eyes were oriented centrally. We subtracted No probe trials from Congruent and Incongruent trials to yield visual evoked potentials (VEPs) free from the influence of prior somatosensory processing. Comparably to prior findings in adults, 8-month-olds' first negative component of the VEP was enhanced in the hemisphere ipsilateral to the probe when the visual probe was spatially congruent with the tactile cue. No crossmodal effect was observed in the 5-month-olds, indicating developments in the crossmodal coordination of attention in the first year of life. ### Competing Interest Statement The authors have declared no competing interest. European Research Council, https://ror.org/0472cxd90, FP7/2007-2013 Leverhulme Trust, https://ror.org/012mzw131, ECF-2019-563

Here, we demonstrate the critical role of developmental vision in the origins of a neural processing hierarchy in which somatosensory events are mapped from somatotopic locations onto a body representation with respect to external space. Both sighted and severely visually impaired infants showed prominent contralateral somatosensory activation in uncrossed- and crossed-hands postures in the early feedforward stages of processing (45-65 ms). By 105-120 ms following the somatosensory stimulus, anatomically ipsilateral activation was observed in the crossed-hands posture in sighted infants only, which was directly related to behavioral orienting toward the incorrect hand, reflecting a remapping of touches with reference to the position of the limbs in external space. Severely visually impaired infants exhibited a somatosensory response that was persistent contralateral across early and later processing stages, with the early responses associated with behavior. These findings demonstrate how visual experience in early postnatal development constructs the neural-behavioral basis of embodied spatial perception.

There is a clue in the name. 'Infant' is derived from the Latin in fans (without speech). Human babies cannot report their experiences and are uncooperative (to say the least) when it comes to experimental task instructions. For these reasons, it has been difficult to establish when babies become conscious. Bayne and colleagues propose a cluster-based methodology for overcoming these issues, arguing that consciousness emerges in the last prenatal trimester [ 1. Bayne T. et al. Consciousness in the cradle: on the emergence of infant experience. Trends Cogn. Sci. 2023; 27: 1135-1149 Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar ]. We are heartily enthusiastic about this approach but consider some complications. While Bayne et al. identify behavioural and neural markers of consciousness commensurate with an 'early emergence' view, we note that other markers point to a 'late emergence' view. In the spirit of optimism, we suggest how the cluster-based methodology may overcome this problem (Box 1). Box 1Let the data choose!Some markers of consciousness emerge at an early stage, others at a later stage. To establish when consciousness emerges, we need to distinguish the reliable markers (that correctly indicate consciousness) from misleading ones. The solution is correlations between markers. If a potential marker co-occurs with other markers within a given age group, this raises our confidence that those markers reliably indicate consciousness and its emergence. Similarly, we can identify a cluster of reliable markers if the strength of a particular marker at an early development stage (e.g., at birth) predicts longitudinally in individual infants the strength or earlier emergence of another marker at a later stage (e.g., at 6 months of age). If a potential marker does not co-occur with or predict others, this raises confidence that it is misleading. This presents a data-driven methodology, to sift reliable from misleading markers and establish the onset of consciousness in early life. The availability of multivariate pattern analysis which can be applied across brain and behaviour, even between infants and adults [ 8. Spriet C. et al. Visual object categorization in infancy. Proc. Natl. Acad. Sci. U. S. A. 2022; 119e2105866119 Crossref PubMed Scopus (14) Google Scholar ], and longitudinal modelling approaches (e.g., [ 9. Kievit R.A. et al. Developmental cognitive neuroscience using latent change score models: a tutorial and applications. Dev. Cogn. Neurosci. 2018; 33: 99-117 Crossref PubMed Scopus (238) Google Scholar , 10. Devine R.T. Hughes C. Relations between false belief understanding and executive function in early childhood: a meta-analysis. Child Dev. 2014; 85: 1777-1794 Crossref PubMed Scopus (302) Google Scholar ]) provides excellent opportunities to capitalize on the potential of cluster correlations. Some markers of consciousness emerge at an early stage, others at a later stage. To establish when consciousness emerges, we need to distinguish the reliable markers (that correctly indicate consciousness) from misleading ones. The solution is correlations between markers. If a potential marker co-occurs with other markers within a given age group, this raises our confidence that those markers reliably indicate consciousness and its emergence. Similarly, we can identify a cluster of reliable markers if the strength of a particular marker at an early development stage (e.g., at birth) predicts longitudinally in individual infants the strength or earlier emergence of another marker at a later stage (e.g., at 6 months of age). If a potential marker does not co-occur with or predict others, this raises confidence that it is misleading. This presents a data-driven methodology, to sift reliable from misleading markers and establish the onset of consciousness in early life. The availability of multivariate pattern analysis which can be applied across brain and behaviour, even between infants and adults [ 8. Spriet C. et al. Visual object categorization in infancy. Proc. Natl. Acad. Sci. U. S. A. 2022; 119e2105866119 Crossref PubMed Scopus (14) Google Scholar ], and longitudinal modelling approaches (e.g., [ 9. Kievit R.A. et al. Developmental cognitive neuroscience using latent change score models: a tutorial and applications. Dev. Cogn. Neurosci. 2018; 33: 99-117 Crossref PubMed Scopus (238) Google Scholar , 10. Devine R.T. Hughes C. Relations between false belief understanding and executive function in early childhood: a meta-analysis. Child Dev. 2014; 85: 1777-1794 Crossref PubMed Scopus (302) Google Scholar ]) provides excellent opportunities to capitalize on the potential of cluster correlations.

This study investigated whether infants encode better the features of a briefly occluded object if its movements are specified simultaneously by vision and audition than if they are not (data collected: 2017-2019). Experiment 1 showed that 10-month-old infants (N = 39, 22 females, White-English) notice changes in the visual pattern on the object irrespective of the stimulation received (spatiotemporally congruent audio-visual stimulation, incongruent stimulation, or visual-only; eta p2$$ {\eta}_{\mathrm{p}}<^>2 $$ = .53). Experiment 2 (N = 72, 36 female) found similar results in 6-month-olds (Test Block 1, eta p2$$ {\eta}_{\mathrm{p}}<^>2 $$ = .13), but not 4-month-olds. Experiment 3 replicated this finding with another group of 6-month-olds (N = 42, 21 females) and showed that congruent stimulation enables infants to detect changes in object trajectory (d = 0.56) in addition to object pattern (d = 1.15), whereas incongruent stimulation hinders performance.

We asked whether, in the first year of life, the infant brain can support the dynamic crossmodal interactions between vision and somatosensation that are required to represent peripersonal space. Infants aged 4 (n = 20, 9 female) and 8 (n = 20, 10 female) months were presented with a visual object that moved towards their body or receded away from it. This was presented in the bottom half of the screen and not fixated upon by the infants, who were instead focusing on an attention getter at the top of the screen. The visual moving object then disappeared and was followed by a vibrotactile stimulus occurring later in time and in a different location in space (on their hands). The 4-month-olds’ somatosensory evoked potentials (SEPs) were enhanced when tactile stimuli were preceded by unattended approaching visual motion, demonstrating that the dynamic visual-somatosensory cortical interactions underpinning representations of the body and peripersonal space begin early in the first year of life. Within the 8-month-olds’ sample, SEPs were increasingly enhanced by (unexpected) tactile stimuli following receding visual motion as age in days increased, demonstrating changes in the neural underpinnings of the representations of peripersonal space across the first year of life.

Human infants cannot report their experiences, limiting what we can learn about their bodily awareness. However, visual cortical responses to the body, linked to visual awareness and selective attention in adults, can be easily measured in infants and provide a promising marker of bodily awareness in early life. We presented 4- and 8-month-old infants with a flickering (7.5 Hz) video of a hand being stroked and recorded steady-state visual evoked potentials (SSVEPs). In half of the trials, the infants also received tactile stroking synchronously with visual stroking. The 8-month-old, but not the 4-month-old infants, showed a significant enhancement of SSVEP responses when they received tactile stimulation concurrent with the visually observed stroking. Follow-up experiments showed that this enhancement did not occur when the visual hand was presented in an incompatible posture with the infant’s own body or when the visual stimulus was a body-irrelevant video. Our findings provide a novel insight into the development of bodily self-awareness in the first year of life.

Knowledge of one's own body size is a crucial facet of body representation, both for acting on the environment and perhaps also for constraining body ownership. However, representations of body size may be somewhat plastic, particularly to allow for physical growth in childhood. Here we report a developmental investigation into the role of hand size in body representation (the sense of body ownership, perception of hand position, and perception of own-hand size). Using the rubber hand illusion paradigm, this study used different fake hand sizes (60%, 80%, 100%, 120% or 140% of typical size) in three age groups (6- to 7-year-olds, 12- to 13-year-olds, and adults; N = 229). We found no evidence that hand size constrains ownership or position: participants embodied hands which were both larger and smaller than their own, and indeed judged their own hands to have changed size following the illusion. Children and adolescents embodied the fake hands more than adults, with a greater tendency to feel their own hand had changed size. Adolescents were particularly sensitive to multisensory information. In sum, we found substantial plasticity in the representation of own-body size, with partial support for the hypothesis that children have looser representations than adults.

Adults’ body representation is constrained by multisensory information and knowledge of the body such as its possible postures. This study (N = 180) tested for similar constraints in children. Using the rubber hand illusion with adults and 6- to 7-year-olds, we measured proprioceptive drift (an index of hand localisation) and ratings of felt hand ownership. The fake hand was either congruent or incongruent with the participant’s own. Across ages, congruency of posture and visual-tactile congruency yielded greater drift towards the fake hand. Ownership ratings were higher with congruent visual-tactile information, but unaffected by posture. Posture constrains body representation similarly in children and adults, suggesting that children have sensitive, robust mechanisms for maintaining a sense of bodily self.

Children’s and adults’ body representation is constrained by bottom-up multisensory information and by top-down knowledge on possible postures. Using the rubber hand illusion paradigm, this study (N = 229) investigates whether different fake hand sizes (60%, 80%, 100%, 120% or 140% of typical hand size) constrain embodiment in three age groups (6- to 7-year-olds, 12- to 13-year-olds, and adults). Embodiment was measured by questionnaire, proprioceptive drift, and affordance judgements. In line with previous work, we found robust effects of age and synchrony, with higher responses at younger ages and under conditions of visual-tactile synchrony. There were no significant effects of hand size on proprioceptive drift or self-rated hand ownership; nor did participants verbally report that their hand had changed size. Participants of all ages therefore embodied a differently-sized fake hand, without being explicitly aware of the size change. However, manual judgments of own-hand size were significantly influenced by the size of the previously seen fake hand. Therefore, participants did implicitly incorporate a size change into their body schema. In sum, embodiment of differently-sized hands reveals substantial plasticity in body representation, modulated strongly by multisensory information and age. Further, the embodiment of a differently-sized hand specifically affects action-oriented representations of the body.

Perceiving one’s own body underpins skilled interactions with the external world and plays a fundamental role in the sense of self. Findings across experimental psychology and neuroscience show that body perception depends on integrating bodily information across multiple senses. However, the emergence of such multisensory abilities in early human development is just starting to be investigated. It is now generally established that human infants are sensitive to the spatiotemporal congruency between cues about the body coming from different senses, even with only a few months or even days of postnatal experience. Conversely, other abilities appear to have a more protracted development, such as the ability to make the crossmodal links required to locate tactile stimuli in external space, i.e., the “remapping problem” (Driver & Spence, 1998; Heed, Buchholz, Engel, & Röder, 2015), which develops gradually in the first year of life. This article briefly reviews the scientific literature concerning body representations in early infancy, highlighting the important role of visual experience in the development of these fundamental representational abilities.

The ability to resist distracting stimuli whilst voluntarily focusing on a task is fundamental to our everyday cognitive functioning. Here, we investigated how this ability develops, and thereafter declines, across the life-span using a single task/experiment. Young children (5?7 years), older children (10?11 years), young adults (20?27 years), and older adults (62?86 years) were presented with complex visual scenes. Endogenous (voluntary) attention was engaged by having the participants search for a visual target presented on either the left or right side of the display. The onset of the visual scenes was preceded ? at stimulus onset asynchronies (SOAs) of 50, 200, or 500 ms ? by a task-irrelevant sound (an exogenous crossmodal spatial distractor) delivered either on the same or opposite side as the visual target, or simultaneously on both sides (cued, uncued, or neutral trials, respectively). Age-related differences were revealed, especially in the extreme age-groups, which showed a greater impact of crossmodal spatial distractors. Young children were highly susceptible to exogenous spatial distraction at the shortest SOA (50 ms), whereas older adults were distracted at all SOAs, showing significant exogenous capture effects during the visual search task. By contrast, older children and young adults? search performance was not significantly affected by crossmodal spatial distraction. Overall, these findings present a detailed picture of the developmental trajectory of endogenous resistance to crossmodal spatial distraction from childhood to old age and demonstrate a different efficiency in coping with distraction across the four age-groups studied.

Adults’ body representation is constrained by multisensory information and knowledge of the body such as its possible postures. This study ( N  = 180) tested for similar constraints in children. Using the rubber hand illusion with adults and 6- to 7-year olds, we measured proprioceptive drift (an index of hand localization) and ratings of felt hand ownership. The fake hand was either congruent or incongruent with the participant’s own. Across ages, congruency of posture and visual–tactile congruency yielded greater drift toward the fake hand. Ownership ratings were higher with congruent visual–tactile information, but unaffected by posture. Posture constrains body representation similarly in children and adults, suggesting that children have sensitive, robust mechanisms for maintaining a sense of bodily self.

Data from the study "Multisensory spatial perception in visually impaired infants". Data are in textual tab-delimited format. Summary Congenitally blind infants are not only deprived of visual input but also of visual influences on the intact senses. The important role that vision plays in the early development of multisensory spatial perception1, 2, 3, 4, 5, 6, 7 (e.g., in crossmodal calibration8, 9, 10 and in the formation of multisensory spatial representations of the body and the world1,2) raises the possibility that impairments in spatial perception are at the heart of the wide range of difficulties that visually impaired infants show across spatial,8, 9, 10, 11, 12 motor,13, 14, 15, 16, 17 and social domains.8,18,19 But investigations of early development are needed to clarify how visually impaired infants’ spatial hearing and touch support their emerging ability to make sense of their body and the outside world. We compared sighted (S) and severely visually impaired (SVI) infants’ responses to auditory and tactile stimuli presented on their hands. No statistically reliable differences in the direction or latency of responses to auditory stimuli emerged, but significant group differences emerged in responses to tactile and audiotactile stimuli. The visually impaired infants showed attenuated audiotactile spatial integration and interference, weighted more tactile than auditory cues when the two were presented in conflict, and showed a more limited influence of representations of the external layout of the body on tactile spatial perception.20 These findings uncover a distinct phenotype of multisensory spatial perception in early postnatal visual deprivation. Importantly, evidence of audiotactile spatial integration in visually impaired infants, albeit to a lesser degree than in sighted infants, signals the potential of multisensory rehabilitation methods in early development. Orienting responses and reaction times (RT) are reported, based on the scoring of two independent naive raters, for each trial of each subject, group (SVI/S), posture (Uncrossed/Crossed), and sensory condition (Tactile only, Auditory only, Audiotactile congruent, Audiotactile incongruent). Trial is the trial number, condition is the sensory condition, audio and tactile respectively refer to the side of the stimulated hand, response_status reports if the response is defined or undefined, response modality reports if the modality used by subjects to respond/not to respond to stimuli (hand, eye, both hands, no motion), group is if the subject was a sighted (S) or a severely visually impaired (SVI) infant, age_mounth is the age expressed in months, RT_rater1, RT_rater 2 and RT are respectively the RT assigned by the two raters and the merge of the two estimations (for RTs, the mean), the same organization for response_side, and for response_modality (for those variables, when the estimation of the two raters did not agree, the merged classification was set to unknown, that is uncertain/undefined).

Two experiments examined perceptual colocation of visual and tactile stimuli in young infants. Experiment 1 compared 4- ( n  = 15) and 6-month-old ( n  = 12) infants’ visual preferences for visual-tactile stimulus pairs presented across the same or different feet. The 4- and 6-month-olds showed, respectively, preferences for colocated and noncolocated conditions, demonstrating sensitivity to visual-tactile colocation on their feet. This extends previous findings of visual-tactile perceptual colocation on the hands in older infants. Control conditions excluded the possibility that both 6- (Experiment 1), and 4-month-olds (Experiment 2, n  = 12) perceived colocation on the basis of an undifferentiated supramodal coding of spatial distance between stimuli. Bimodal perception of visual-tactile colocation is available by 4 months of age, that is, prior to the development of skilled reaching.

Congenitally blind infants develop in a perceptual environment which not only lacks visual input but is also deprived of the influences of vision on the intact senses such as in crossmodal calibration and the mapping of the senses to external spatial coordinates. Comparing early spatial representations in blind and sighted infants, we elicited manual orienting responses to auditory and tactile stimuli presented on infants’ hands, manipulating arm posture to examine the frames of reference used for localisation. The sighted infants oriented to auditory cues in audiotactile presentations, and used an external spatial code when locating touches (they showed a crossed hands deficit [6]). The blind infants oriented to tactile cues in audiotactile presentations and used an anatomical spatial code for locating touches, notably showing improved tactile localisation than the sighted participants in the crossed hands condition. Both the sighted and the blind infants demonstrated improved localisation when auditory stimuli were presented at the same time and on the same hand as tactile stimuli, but the blind infants showed a weaker multisensory gain. The blind infants’ responses to touches at the expense of auditory cues indicates an impairment of auditory localisation, leading to a spatial bias to anatomical rather than external coordinates. The finding of multisensory gain across blind and sighted infants demonstrates multisensory integration. This shows that spatial localisation can be enhanced in blind infants by providing spatially congruent stimulation across audition and touch, a finding which bears important implications for approaches to early intervention.

Two experiments examined perceptual colocation of visual and tactile stimuli in young infants. Experiment 1 compared 4- (n = 15) and 6-month-old (n = 12) infants' visual preferences for visual-tactile stimulus pairs presented across the same or different feet. The 4- and 6-month-olds showed, respectively, preferences for colocated and noncolocated conditions, demonstrating sensitivity to visual-tactile colocation on their feet. This extends previous findings of visual-tactile perceptual colocation on the hands in older infants. Control conditions excluded the possibility that both 6- (Experiment 1), and 4-month-olds (Experiment 2, n = 12) perceived colocation on the basis of an undifferentiated supramodal coding of spatial distance between stimuli. Bimodal perception of visual-tactile colocation is available by 4 months of age, that is, prior to the development of skilled reaching.