Neurobiological mechanisms and clinical implications of scene perception impairment in schizophrenia

Abstract

As more recognition has been given to scene perception impairments in schizophrenia as a manifestation of abnormal sensorimotor function leading to subsequent social and cognitive decline. Synthesise the findings of event-related potential, functional magnetic resonance imaging (MRI) scanning, structural MRI, and molecular neuroscience techniques. A meta-analysis demonstrated that there was a moderate-to-large decrease in the visual mismatch negativity (g = -0.63), indicating impairments of the automatic prediction-error signal; while abnormal N170 and late positive potentials suggested deficits spreading from early visual analysis to face-selective and socio-affective processing. The structures of the MRIs showed changes in gyrolithogenesis and others to support an objective measurement system for structural-functional relationship. At the mechanical level, N-methyl-D-aspartate receptor hypo-function, parvalbumin interneuron dysfunction, gamma-band disordering, and glial-related neuro-inflammation all affect predictive coding and obtain control together. In practice, these malformations may provide a reason for the inconsistency in the formation of social scenes among some patients when registering partial visual information. Accordingly, we propose that scenes of perception should not be regarded merely as minor visual issues but rather as a clinically significant system-level objective. Visual remediation, neuro-modulatory, ecological evaluation, biomarker-guided intervention have emerged as particularly relevant, but long-term and mechanisms-supported clinical studies are lacking.

Key Words: Schizophrenia; Scene perception; Electrophysiology; Structural magnetic resonance imaging; Effective connectivity; Predictive coding

Core Tip: Scene perception impairment in schizophrenia is best understood as a multilevel disturbance that begins with early sensory prediction deficits and unfolds through visual-stream dysfunction, network dysconnectivity, and clinically relevant failures in social-scene interpretation. This review incorporates structural magnetic resonance imaging findings, expands the analysis of effective connectivity and feedforward-feedback abnormalities, links molecular evidence across gene, cell, circuit, and computational levels, and sharpens the discussion of translational implications. Taken together, the literature suggests that scene perception may offer a practical bridge between basic visual neuroscience and real-world functional outcomes.

INTRODUCTION

A common type of severe mental disorder characterised by psychotic symptoms and impaired cognition. Schizophrenia has a serious public health problem; however, it may not be fully expressed in terms of hallucinations and delusions[1,2]. Also, many patients have problems with rapid and simple handling of complicated situation perception tasks. Although it receives less attention in the routine clinical presentations; still counts. Successful operation is not limited to isolating objects individually in daily life. To extract the essence of scenes, integrate context information, track social factors, and update predictions in response to environmental changes[3-5].

Therefore, scene perception provides an effective way to enter schizophrenia; that is to say, they are related to initial visual processing social cognition and life adaptability. Practically speaking, a patient might notice some parts of a scene while having trouble interpreting the whole picture: (1) There is a person whose expression appears uncertain; (2) There are movements on display that need to be interpreted correctly; and (3) Context exists here but fails to regulate this phenomenon effectively. Recent meta-analysis study papers focusing on face recognition and related issues concerning the socio-psychological functions have shown that they are merely individual deviations, lacking a systemic perspective. They may instead help shape social withdrawal, misattribution, and functional disability[4,5].

Given that there are more applications, the problem mentioned above is a valid one. Early discussions about schizophrenia generally treated changes in visual perception, social cognition and neurological circuits separately. It has provided some useful information but scatteredly. Furthermore, there is still no unifying explanation at the level of integration among low-level sense disorder, high-scene understanding capability, structural brain abnormalities, and molecular malfunctions. Now that the available evidence has made this integration less unrealistic than before[3,6-15].

Scene perception is the core mechanism of daily adaptation. The current article refers to this term broadly and systematically, encompassing rapid scene gist extraction, context adjustment, facial-motor-based social-cue interpretation, as well as active sampling based on eye movement analysis. Although these processes can be analysed separately, in reality, they work together closely and interact continuously. This part is mainly for patients with schizophrenia. There is no obvious disorder in a single sense of perception. It also interferes with changes among different stages of processing, such as shifting from sensory information reception to perceptual arrangement; from then on, perceive organized social meanings; finally, trigger an adaptive reaction based on these. In this more comprehensive sense, why scenes that have affected perceptual abilities to a greater degree result in broader consequences are also presented.

Two goals of the current article are as follows: Firstly, by combining the current evidence of scene-perception impairment in electrophysiology, functional neuroimaging, structural neuroimaging, molecular-cellular neuroscience and new intervention research. Secondly, a more definite hierarchy exists concerning this problem explicitly. Our working view is that scene-perception impairment in schizophrenia is best conceptualized as a hierarchical systems disturbance, with early prediction and gain-control abnormalities propagating into visual-stream dysfunction, large-scale network dysconnectivity, and clinically meaningful failures of social-scene interpretation. Although this view is somewhat vague, it might help us rather than just a deficiency checklist in linking mechanisms with performance, especially suggesting interventions.

ELECTROPHYSIOLOGICAL EVIDENCE: FROM EARLY CHANGE DETECTION TO SOCIAL-SCENE ENCODING

Because electrophysiological research can determine at which point scene perception begins deviating from the norm. It’s time-consuming now. Recently, a meta-analysis on visual-mismatch-negative-events showed that the average reduction rate for schizophrenia patients was as high as 0.63 (g = -0.63), which supports poor automatic correction ability and reduced prediction error signal transmission during early stages[6]. That is to say, the brain cannot register as much that the visual world violates its own expectations; as the disorder’s discussion shifts from predictive coding to another level, it cannot be considered trivial at all[10].

The evidence of face-and social-perceptual event-related potential shifts the disturbance up the chain. Abnormal N170 responses, together with altered late positive potential amplitudes during emotional-face processing, indicate that schizophrenia affects both structural encoding of faces and later evaluative integration[16,17]. These forms are notable; it is possible that the scene-perspective issue may arise even if there is a decline in basic vision functions. Instead, early sensory inefficiency and later social-cognitive instability seem to coexist and likely interact.

Older visual electrophysiology is still applicable here. Butler et al[18] described early-stage visual processing and cortical amplification deficits consistent with impaired magnocellular or gain-related mechanisms. More recent contextual modulation studies also support this view by pointing out abnormalities in contrast modulation and decreased utilisation of surrounding context; thus, it can be inferred that the early sensory representation is noisy, inefficiently weighted, or poorly stabilised through local gain regulation[19]. Because scenes contain richer contexts than other parts of the image. Given that the system cannot utilise context on this side, there is no guarantee for the subsequent scenario interpretations.

Taken together, the event-related potential literature suggests a staged disturbance rather than a single lesion. Automatic deviation detection weakens; face-sensitive processing efficiency decreases; later the social emotional assessment is also disrupted[6,16-19]. Moreover, the evidence has clear limits: Much of the literature relies on facial-processing, emotion-processing, or contextual-modulation paradigms rather than full naturalistic scenes. Most studies are cross-sectional and cannot distinguish between trait vulnerability, disease stage effect, or sensitivity to treatment[6,16-19]. However, the convergent temporal pattern can be used to treat scene perception as multiple components of targets rather than just visual symptom clusters. Table 1 outlines an overall structure for presenting primary electrophysiological indices covered herein: Early change-detection deficits, face-sensitive discrepancies, and impaired active visual probing are among them.

Table 1 Electrophysiological markers associated with impaired scene perception in schizophrenia.

Ref.	Task domain	Main findings	Explanatory value	Level of evidence
[6]	VMMN	Meta-analysis showed reduced VMMN amplitude (g = -0.63), consistent with impaired automatic visual prediction error signal	Support for early predictive coding interference in scene change detection	Meta-analysis
[16,17]	The role of N170 and LPP in face/emotion processing	Abnormal N170 and reduced late positive responses indicate lower face encoding efficiency and later social-emotional integration	It is suggested that the scene perception deficit extends from face-sensitive processing to the evaluation phase	Event-related potentials in humans
[18,19]	Early visual processing/environmental regulation	The background of the early visual magnification and abnormal dependent contrast modulation means less gain control	4 associate low levels of sensory instability with later scene integration difficulties	Research on human mechanisms
[35,36]	Eye movement and eye fixation-related signals	Changes in fixation-related saccades and abnormalities in fixation-related potentials can be observed during natural fixation	Emphasis on active visual sampling as part of the clinically relevant scene perception	Naturalistic human research

VMMN: Visual mismatch negative wave; LPP: Late positive potential.

NEUROIMAGING EVIDENCE: HIERARCHICAL VISUAL STREAMS, EFFECTIVE CONNECTIVITY, AND ECOLOGICAL NETWORK DYNAMICS

Functional neuroimaging adds spatial elements to the time course. The existing task-based research shows that there are defects at multiple levels of vision formation: First layer cortex, then upper-order occipitotemporal region and lower-order temporal-parietal part[3,7,19-21]. Therefore, the scene-perception disturbance in patients with schizophrenia should not be considered a single lesioned site. Better regarded as an interruption sequence from the perception of sense organs to object identification, face detection, etc., followed by motion recognition and ultimately social context judgment.

At the bottom of this hierarchical system, the primary visual cortex is neither uniformly absent from activity nor uniformly hyporeceptive. It is less explicit. Contextual-modulation studies indicate that V1 responses can be atypical under conditions that normally sharpen figure-ground segregation and spatial attention[19]. This fits a gain-control account: The issue may not be simple underactivation but unstable tuning. When there is an imbalance, the forward feedback at the mid- and high-level visuomotor areas would lose more information.

At a higher level, convergence of tasks is focused on fusiform gyri, middle temporal visual area complex (MT+), posterior superior temporal sulcus (pSTS) and lateral occipital areas. The most relevant among the current “third-visual-pathway” theories is that this link between motion sensitivity and social awareness differs from the traditional concept of two independent domains[3]. Schizophrenia shows a reduction in activity levels of the MT+ and pSTS when performing an observed or simulated movement task; thus, there is impairment in dynamically interpreting social information here. The face is no longer to be perceived but embedded within a sequence of scenes. It is not robust.

There are few direct evidences of feedforward and feedback signaling in patients with schizophrenia under scenesensing task. Although these existing results still cannot conclusively indicate that disturbances are limited to specific areas of hypo- or hyperactivity but rather extend to abnormal top-down regulation in multiple distributed network areas. Blain et al[7] applied dynamic causal models during gaze processing; they found abnormal effective connectivity between the dorsomedial prefrontal cortex and pSTS and inferior parietal regions, which was correlated with symptom severity and social adaptation. Gaze processing, in other words, is merely one element of scene understanding and does not replace the latter entirely. Nevertheless, this study offers some enlightenment; that is to say, there may be problems with the transmission of information for people who have schizophrenia. Conveyed through a relevant message in dynamic connection studies of the lateral occipital cortex. Li et al[20] found that the connectivity of latera occipital cortex between resting state and task states was abnormal in schizophrenic patients; therefore, it indicated poor functional flexibility for spatial perception switching. As such, this is an important improvement; in the field of scene understanding, flexibility among several basic processes needs to be realised online. If, under strict spatial limitations on location-based interactive information or its environmental adaptability problems exist that affect the consistency of scene perception among people.

The naturalistic paradigm reinforces this interpretation. During continuous movie-like stimulation, patients with first-episode psychosis already show altered functional network connectivity and topology[21]. These models are not replacing the traditional tasks but extending them. The conventional event-related design can separate each part distinctly. However, naturalistic paradigms are closer to the actual computation required by scenes; multiple cues evolve simultaneously; attention shifts continuously; and internal models need to be recalibrated without clear trial markers. Therefore, from this perspective, naturalistic functional magnetic resonance imaging (MRI) can be considered particularly suitable as a reference for examining the scene perception phenomenon; that is, by simulating actual scenarios where patients may face difficulties at present.

Complete the situation of large-scale networks. Across studies, there is a decline in effective cooperation among inputting visual information and selecting internally based on the dysfunction of the anterior cingulate, default mode, salience and executive-control networks[22-25]. Some of these deviations are probably trans-diagnostic; however, for patients with schizophrenia, it often acquires a specific clinical significance as part of the abnormal sense-of-self disturbance. Although there is still no direct evidence of scene-specific feedforward and feedback signalization; most existing arguments are based on gaze, motion, or context modulation or hub-based proxy paradigm, which cannot fully reflect whole-scene tasks[3,7,19-25]. The naturalistic paradigm improves ecological feasibility but still cannot address the question of where abnormalities in the information transmission map to particular scenes’ sub-elements. Thus, the literature supports a hierarchical dysconnectivity account but not yet a fully unified scene-specific model. Figure 1 presents a high-level organisation at the system level for scene perception; the main visual pathways and brain networks involved in schizophrenia are shown to illustrate their hierarchical structure, including disturbed connections among damaged parts. The following comprehensive summary table was also prepared based on the function-related neuroimaging findings reported above: Hierarchical visual-system abnormalities; disrupted functional connectivity; naturalistic network dysfunction (Table 2).

Figure 1 A schematic overview of visual pathways and brain networks associated with scene perception in schizophrenia. This figure summarizes the main visual pathways and high-order brain networks associated with scene perception in schizophrenia, illustrating the hierarchical processing of visual information from early processing in the primary visual cortex to the ventral, dorsal and third visual pathways, as well as the interactions between these pathways and large-scale brain systems responsible for network integration and cognitive control. This figure highlights the potential pathological basis of scene perception impairment in schizophrenia: Abnormalities in information flow, effective connectivity and network coordination. LOC: Lateral occipital cortex; dmPFC: Dorso-medial prefrontal cortex; ACC: Anterior cingulate cortex; DMN: Default mode network; MT+: Middle temporal visual area complex; pSTS: Posterior superior temporal sulcus.

Table 2 Functional neuroimaging performance: Hierarchical visual processing, effective connection and naturalistic paradigm.

Ref.	Imaging focus	Main findings	Relevance to scene perception	Level of evidence
[3]	Third visual pathway/MT+/pSTS	Real and implied movement of activation means less sensitive to exercise social perception and processing	Disturbances in support of the process of translating dynamic visual input into socially meaningful scene information	Task functional magnetic resonance imaging
[7]	Dynamic causal models during gaze processing	Abnormal top-down effective connectivity from the dmPFC to the pSTS and inferior parietal regions has been linked to symptoms and social functioning	Provides direct evidence that feedforward feedback coordination is disrupted	Effective connectivity research
[20]	Lateral occipital cortex dynamic connectivity	LOC connectivity is abnormally reconfigured during resting and task states	During the prompt changes in the demand of the scene, unstable coordinate object processing hub	Dynamic connectivity research
[21]	Natural fMRI of first-episode psychosis	Movie-like stimulation reveals alterations in whole-brain connectivity and network topology in the early stages of psychosis	Improve ecological validity and complement task-based findings	Natural fMRI
[22-25]	ACC, DMN, salience, executive control, and coupling studies	Large-scale connectivity impairments reduce coordination between sensory input, salience allocation, and internal guidance control	This explains why scene interpretation fails even when some visual information is retained	Network-level human body imaging

MT+: Middle temporal visual area complex; pSTS: Posterior temporal channel; dmPFC: Dorso-medial prefrontal cortex; LOC: The lateral occipital cortex; fMRI: Functional magnetic resonance imaging; ACC: Anterior cingulate cortex; DMN: Default mode network.

STRUCTURAL MRI FOUNDATIONS OF SCENE-PERCEPTION IMPAIRMENT

Structural MRI results are worth considering separately at present. Functional disorders often do not stand alone with a dysfunction in bodily functions; rather, we should pay attention to complex perceptual movements and so on. Although structural images on their own cannot explain scenes-perception disorder; lateral anatomical outline will point out where this anomaly occurs. There has been a reduction in grey matter, cortical thinning or other abnormalities at the level of the visual-analytical area that make it difficult for subjects to acquire, process, store sensory information via an extra-demanding higher-order pathway.

Several lines of evidence are connected. Schultz et al[8] reported altered cortical shape in the visual cortex, with increased gyrification in V1, V2, and MT+ and thinning in MT+, providing early in vivo support for disturbed visual system morphology in patients with schizophrenia. Two factors of these results are noteworthy. They put morphological changes directly inside the visual region rather than only in the frontotemporal associative area. Secondly, in conjunction with hypergyria and a selective thinned cortex, it suggests both neurodevelopmental influences and later development are related to this.

When structure-function relationships are valuable because of anatomic variations leading to functional defects, etc. Turetsky et al[26] found that impaired face-based emotional perception among patients with schizophrenia exhibited a decrease in the N170 response; moreover, this decrement was closely related to reduced fusiform grey matter size. This result cannot provide an obvious causality relationship; however, the gap from shape to signal dysfunction is reduced somewhat. That is, when the fusiform structure is damaged, there may be weakened face-sensitive electrophysiological responses; that is, as brain anatomy changes, it can reflect perception disorders to some extent.

Broader morphometric studies reinforce this point while also reminding us to stay cautious. Sasabayashi et al[27] described increased gyrification across the schizophrenia spectrum, a pattern often interpreted as a marker of early neurodevelopmental deviation. García-León et al[28] further reported regional cortical volume abnormalities associated with symptoms and cognitive impairment. These studies are not scene-specific in a narrow experimental sense; however, they remain relevant because scene perception involves the same occipito-temporal and associative systems whose structure is altered across the disorder.

Regarding white matter structure as well. The structural-functional abnormality of callosal-white matter-cortical circuitries suggests a disruption to the interhemispheric integration in schizophrenia patients[22]. In scene perception under a naturalistic environment, it needs to integrate multiple kinds of information quickly. Therefore, reduced integrity of the bridge-bridging pathway may further exacerbate local visual abnormalities through decreased overall coordination efficiency.

To sum up, at present, there are limitations in some parts of the research results from current structural MRI literature: Direct scene-specific morphometric studies are relatively rare; some arguments depend on neighboring fields such as face recognition, motion perception, and general vision system structure[8,22,26-28]. Therefore, the structure-function relationship can be understood as being restricted but not finalised. Despite this, there is enough accumulated evidence to make a defence of pure functionality improbable. There is a better interpretation of this scene-perception disorder as being caused by structural-function cascades; alterations to cortical folding, thickness, volume, and white matter organisation affect boundary formations during compensatory functioning in patients with schizophrenia. The systematic summarization of the structural and white matter MRI evidence, which includes gyrification abnormality, reduction in cortical thickness or volume, impairment of white matter connectivity; structure-function association (Table 3).

Table 3 Structural and white matter magnetic resonance imaging evidence associated with impaired scene perception in schizophrenia.

Ref.	Structural MRI domains	Key findings	The structure and meaning	Level of evidence
[8]	Visual cortex retraction/thickness	V1, V2, and the MT+ area increase, the MT+ area selective thinning	This shows that the visual system forms in the center of the perceptual organization and motion analysis area had changed	Directional structure of MRI
[26]	Gray matter in the fusiform gyrus	Fusiform gray matter decreased with facial emotion recognition disorders; reduced spindle volume was associated with diminished N170 responses	Provides a bridge between anatomy and electrophysiology	Structure research
[27]	Rotational phenomena in the schizophrenia spectrum	There was an increase in gyri across the schizophrenia spectrum, consistent with early neurodevelopmental bias	Support the view that the scene perception of vulnerability may partly reflect the change of cortical development	Spectrum structure of MRI
[28]	Local cortical volume abnormality	Extensive cortical volume differences are associated with symptoms and cognition	It is suggested that deficits in scene perception may be limited by a broader load on occipitotemporal and related structures	Clinical form metrology research
[22]	Corpus callosum white matter organization	Structure - function abnormalities involved hemispheres pathways and visual correlation circuits	Means that the complex scene perception required massive consolidation efficiency reduced	Structure and multimodal MRI

MRI: Magnetic resonance imaging; V1: Primary visual cortex; V2: Secondary visual cortex; MT+: Temporal middle visual area complex.

MOLECULAR CIRCUITRY AND NEUROMODULATORY MECHANISMS

At the molecular level, N-methyl-D-aspartate (NMDA) receptor hypofunction is considered a leading candidate for the mechanisms. Imaging genetics and pharmaco-functional MRI studies have identified a relationship between NMDA receptor and impaired connectivity between the auditory and visual systems in patients with schizophrenia[9]. New cross-diagnostic research on predictive coding supports these key findings. The study describes psychosis as a clinical condition characterized by impaired precision weighting, inaccurate prediction error signaling, and abnormal error transmission[10]. These operations are active during scene perception. The brain must decide which visual signals are reliable, which changes matter, and which prior expectations should be updated.

The computation is rendered real in cells through this transition. NMDA hypofunctionality is particularly harmful to the interneuronal subpopulation expressing parvalbumin (PV), whose function involves regulating the delay of inhibition and repetitive excitation[11,12]. It can be said that there is no separate deficiency in PV-neurons; there has been persistent molecular evidence of their origin[12]. If PV-mediated inhibition weakens, cortical ensembles become less temporally precise. The likely outcome is reduced signal fidelity and poorer coordination across levels of a visual hierarchy.

Gamma oscillations provide the network-level readout of that cellular problem. McNally and McCarley[13] believe that there is an abnormal γ-band phenomenon, through which abnormalities in the system are reflected; it coordinates functional binding at low frequencies while regulating higher-level cognitive functions. Recent studies on resting-state electroencephalography and magnetoencephalography also indicate a perspective of sustained frequency-specific disorders that correspond to impaired integration of the dispersed network[29]. In terms of scene representation, it is clear that during unstable fast-rhythmic coordination, some scenes will be fragmented rather than integrated into an overall picture with social value by the system.

Neuroinflammation can also increase its level of importance and status, be excluded from analysis entirely. Reviews at present highlight that microgliosis and astrocytosis are more dynamic phenomena than passive markers[14,15]. Several pathways for the relevance of microglia-neuron interaction in schizophrenia were summarised by Hartmann et al[14], including abnormal synaptogenesis; inflammatory response; alteration of neuron-glia communication. Laricchiaut, on the other hand, found that there are differences between astigrytes and microligans in influencing thinking and moods through immune-neural detachment[15]. This field is not homogeneous; further elaboration should not be made. However, there has been a gradual increase in suggesting that glial pathology can influence sensory enhancement, synaptic stability, and network plasticity through which matters for visual context-based processing.

Together, these provide empirical support for a multilayered one. Gene-and-receptor-level deviations shift the equilibrium of excitation-inhibition; cell-type-specific perturbations, particularly to PV interneurons and glial cells, disrupt temporal precision and synaptic homeostasis, leading to circuit-level effects such as altered gamma coordination, unbalanced gain regulation, and inefficient predictive learning[9-15,29]. The data at all levels is scattered. As yet there have been no direct investigations demonstrating how these pathways operate at the scene level in patients with schizophrenia; therefore, some parts of the current summary remain conjectural because they progress from receptors and cellular biology to circuit models and finally perceptual outcomes. However, this framework still provides a consistent pathway to link molecular deviations with the frequent behaviour pattern of partial registration, weak context anchoring, and unstable social cognition. Complementary summaries at the level of genes, molecules, cells, circuits, behaviour, etc., that might be associated with scene-perception dysfunction are shown in Table 4. To present a comprehensive understanding of the existing research findings as shown in Figure 2, which constructs a multi-level pathophysiological framework for scene-perception impairment in patients with schizophrenia by integrating molecular-cell abnormalities with electrophysiological effects, vision-system dysfunctions, broad-scale network disruptions, and clinical consequences.

Figure 2 A multilevel pathological model of scene perception impairment in schizophrenia. This figure summarizes a multilevel analytical framework for scene perception impairment in schizophrenia, integrating behavioral phenotypes, electrophysiological abnormalities, visual pathway and large-scale network dysfunction, structural magnetic resonance imaging alterations, as well as molecular and cellular mechanisms. This model elucidates how abnormalities across the aforementioned levels interact synergistically to impair scene understanding capacity and trigger related functional dysfunctions in patients with schizophrenia. Solid arrows indicate hierarchical cascade propagation between levels, while dashed arrows represent cross-level feedback and regulation. This framework represents an integrated model based on existing research evidence, rather than a strictly linear causal sequence. vMMN: Visual mismatch negativity; LPP: Late positive potential; dmPFC: Dorso-medial prefrontal cortex; pSTS: Posterior superior temporal sulcus; ACC: Anterior cingulate cortex; DMN: Default mode network; MT+: Middle temporal visual areacomplex; NMDAR: N-methyl-D-aspartate receptor; PV: Parvalbumin.

Table 4 Potential of scene perception disorders based on gene-molecules-cell-loop-behavior mechanism.

Ref.	Level	Key mechanisms	The impact of speculation on scene perception	Level of evidence
[9-11]	Gene receptor	Connectivity disorders associated with NMDA receptors and NMDA receptor hypofunction alter predictive coding, recurrent gain, and long-term coordination	Impaired efficient updating of accurate weighting and scenario-based predictions	Imaging genetics + mechanistic review
[12,13,29]	Cell oscillations	PV - interneurons defects and abnormal reduces the gamma band suppression time and coordination	Affecting feature binding, context stabilization, and rapid integration of scene elements	Cell-type review + oscillation review
[14,15]	Inflammation of glial cells	Microglia and astrocytes abnormalities may interfere with synaptic pruning, inflammation, signaling communication - glial cells and neurons	May disrupt sensory acquisition control and plasticity associated with situational processing	System evaluation
[10-15]	Circuit calculation	Predictive coding failure, E/I imbalance, and inflammatory regulation converge at the circuit level	Provides a multi-level interpretation of the partial distribution of quasibut unstable scene interpretation	Integrated interpretation
[3,6,16-21,35,36]	Behavior	The scene-sensing phenotype includes impaired gist extraction, weak contextual modulation, abnormal social scene interpretation, and erratic active visual sampling	Defines the observable behavior level and the mechanism of low assumptions involved	Integrated human evidence

NMDA: N-methyl-D-aspartate; PV: Parvalbumin; E/I: Excitation-inhibition.

INTERVENTION STRATEGIES AND CLINICAL IMPLICATIONS

Intervention research remains in its infancy; however, the coherence of the translational logic has increased accordingly. If the scene perception disorder of patients with schizophrenia is a multi-tiered distortion, then intervention strategies will not work in just one direction. Most of the real applications combine sensory-motor rehabilitation techniques with neuro-rehabilitation treatments and ecological testing devices.

Visual reformation is the initial entry point here. Bergson et al[30] have reported an initial positive result from remediation aimed at contrast enhancement; therefore, they believe that some degree of visual disturbance can be plasticized and behaviour-modifiable. Low-level improvement matters in clinical practice only when it transfers upward, such as scene integration, social comprehension, and functional adaptation. It is still an unsolved problem and should therefore be handled accordingly.

Social-cognitive and virtual-reality interventions are promising precisely because they increase ecological validity. Preliminary trials of virtual reality (VR)-based theory-of-mind intervention and VR-assisted social cognition and interaction training suggest that immersive environments may improve engagement and potentially enhance transfer to everyday situations[31,32]. Protocol work on immersive VR-based cognitive remediation further indicates that the field is moving toward more realistic training contexts rather than exclusively laboratory-style exercises[33]. For a topic centered on scene perception, that shift is especially appropriate.

Also under active investigation are biomarker-directed subgroups. Multimodal MRIs and connectomics classifier studies have already demonstrated the ability of individual-level discrimination to some extent[34]. Eye-movement measurements and fixations-saccades deviations are also appealing due to their lower burden and close connection with active observation of natural scene areas[35,36]. One direction for the future might be a combination with electrophysiological examination, eye-trackers etc., for determining injury pathways and providing diagnostic aids or treatment plans collaboratively. We have not yet achieved that level. Still, there are more logical expressions today than a few years ago.

Finally, the ecological approach and the momentaner approach also need to be considered. Based on smartphone- and games-based assessment framework, some research has moved beyond clinics to study people’s behaviour in their everyday lives[37]. There is a deficiency of convenience. Scene perception refers to a kind of ecological function. By testing them separately, it cannot be determined that there will be some important improvement in reality. In other words, the function needs to pay attention to two aspects: Whether a patient recognises a deviation from normal stimulation within the scanning device; whether one can perceive and handle various social interactions in dense environments more accurately. An organised summary of the mechanism-aimed intervention directions, biomarker-driven strategies, and unaddressed translational obstacles is presented in Table 5.

Table 5 Mechanism-based translational research directions and future priority research topics.

Ref.	Directions	Existing evidence	Practical implications	Level of evidence
[30]	Visual restoration	Contrast processing repair improves low-level visual performance in patients with schizophrenia	This is a promising entry point, but the transfer to real-world scene understanding still needs to be demonstrated	Preliminary clinical trials
[31-33]	Based on VR and social cognitive intervention	Early VR studies have shown the feasibility and preliminary efficacy of theory of mind and social cognitive training	It can be used in ecological task design and may be used in scenario-based rehabilitation	Pilot study
[34-36]	Biomarker-guided stratification	Multimodal image classifiers and eye movement markers can distinguish patients and index active visual sampling abnormalities	It may help to identify subgroups and personalize intervention targets	Studies of biomarkers
[37]	Ecological assessment	Smartphones and game-based assessment/intervention approaches are moving beyond the clinic	Encourage measurement of real-world outcomes, not symptom-based endpoints alone	Review points
[10-15,21,22]	Future research priorities	Need longitudinal modal design, better defined scenarios, and stronger structure-function analysis	Will clarify period, reactive mechanism, and therapy	Comprehensive synthesis

VR: Virtual reality.

CONCLUSION

Current evidence supports a fairly consistent conclusion: Scene-perception impairment in patients with schizophrenia is real, clinically meaningful, and mechanistically layered. It is not reducible to a generic attentional problem, nor is it captured by a single visual deficit. The disturbance appears to begin early, with abnormal prediction-error signaling, weak contextual modulation, and impaired sensory gain control[6,10,18,19]. It then propagates through higher-order visual regions and network-level control systems that are needed to assign meaning, social value, and behavioral relevance to complex scenes[3,7,20,21,24,25].

Based on the structure of MRI. Altered gyri-facies, regional thicknesses, fusiform grey matter reductions and callosal-white-matter abnormalities indicate that the scene perception disorder forms in a measurable anatomical background rather than developing exclusively as states-related functional artefacts[8,22,26-28]. Furthermore, the molecular and cellular results are more specific to the pathways leading to local circuit instability causing systems-level perception disorder: NMDA receptor hypofunction, PV interneuron loss, γ-banding disorganisation, and microglia-astrocyte signaling[9-15,29].

Several priorities are now especially urgent. Firstly, in future studies, we will set up more reasonable scene-perception experiments to eliminate this confusion about whether each process can be considered an isolated phase or not. Secondly, longitudinal multi-modality works need to be conducted on patients with clinical high-risk, first-episode, and chronic features to investigate the timing of dysfunction development and predictive biomarkers. Finally, strengthen the structure-function analysis here. Based on the available fields, there exists an uncertain assumption regarding morphological abnormalities affecting electrophysiological phenomena and functions; as a result of this, relatively less specific research has investigated their direct relationship thus far. Intervention studies should go beyond symptom outcomes to assess whether patients are able to navigate visual-and-social spaces better after intervention[29-37].

Then the fundamental problem has nothing to do with abnormally perceived scenes in schizophrenia. Specifically, scene perception may serve as a link to connect sensory neuroscience with systems imaging techniques, computational theory, and various daily-life disabilities in an interdisciplinary way via clinical application. That assertion is still unconfirmed; further scene-specific, long-term tracking data are required. Although at present there are few reports in the field addressing scene perception solely peripherally but rather considering it an important link among neural basis and functional consequences.

ACKNOWLEDGEMENTS

The authors express their gratitude to the Affiliated Mental Health Center of Jiangnan University for its institutional support. Furthermore, we extend our appreciation to the reviewer for their invaluable comments, which significantly enhanced the quality of this paper.