Global neural self-disturbance in schizophrenia: A systematic fMRI review

There is a general consensus that schizophrenia (SZ) is characterized by major changes in the sense of self. Phenomenological studies suggest that these changes in the sense of self stem from a basic disturbance, hence the term ‘basic self-disturbance ’ . While imaging studies demonstrate changes in various regions during self-focused tasks, the exact neural correlates of such basic self-disturbances remain unclear. If the self-disturbance is indeed basic and thereby underlies all other symptoms, one would expect it to be related to more global rather than local changes in the brain. Testing this hypothesis, we conducted a systematic review of fMRI studies on self in SZ. Our main findings are 1. Abnormal activity related to the self can be observed in a variety of different regions ranging from higher-order transmodal to lower-order unimodal regions, 2. These findings hold true across different tasks including self-reflection, self-referentiality, and self-agency, and 3. The global neural abnormalities related to the self in SZ correspond to all layers of the self, predominantly the mental and exteroceptive self. Such global neural disturbance of self converges well with the basic self-disturbance as described in phenomenology.


Introduction
Self-disturbance is a core feature of Schizophrenia (SZ) that can conceptualize a wide range of symptoms, both positive and negative (Henriksen and Parnas, 2012;Sandsten et al., 2022;Sass et al., 2018;Sass and Parnas, 2003).This is what makes self-disturbance a high-yield feature to elicit on neuroimaging in SZ patients.In the current literature, SZ is largely considered a disorder of interconnectivity between brain regions, termed "functional connectivity," studied via resting state fMRI (rs-fMRI) (Menon et al., 2023).In light of this well-researched finding, there is a current shift away from studying SZ symptomology within the context of individual brain areas (Menon et al., 2023).While rs-fMRI and functional connectivity studies may be popular due to minimal patient compliance and lack of task-dependent paradigms, several studies reveal conflicting or inconsistent results when applied to SZ (Cai et al., 2022).Interestingly, task-based neuroimaging studies that elicit specific brain regions come with the advantage of providing a direct link between fMRI findings and psychological constructs (Potvin et al., 2019).
In a first-of-its-kind meta-analysis by Potvin et al., individual brain regions implicated in self-related task-based neuroimaging studies were analyzed.This was done to test the validity of a well-established finding that a group of structures in the cortical midline of the brain referred to as the Default Mode Network (DMN) is implicated in SZ psychopathology -a finding largely derived from rs-fMRI studies (Hu et al., 2017;Liemburg et al., 2012;Martin et al., 2023;Roig-Herrero et al., 2022;van Buuren et al., 2012).The DMN involves areas of the brain like the medial prefrontal cortex (mPFC), anterior cingulate cortex (ACC), and posterior cingulate cortex (PCC) that are cortical midline structures and are active at rest.At rest, the mind is wandering and in introspection, therefore the DMN is hypothesized to be the main network involved in self-referential thinking (Potvin et al., 2019;Raichle, 2015).Potvin et al. found that while they were able to prove the hypothesis partially, they found that regions outside the DMN, particularly the thalamus, were implicated in the psychopathology of SZ, which may point to a more global disturbance.They also appreciate that areas of the cortex including temporal and insular regions are activated in SZ with an unclear significance.This observation is consistent with a recent review showing global rather than local temporal electrophysiological changes in EEG during selfrelated tasks in SZ (Hamilton and Northoff, 2021).This calls into question whether self-disturbance in SZ is mediated by various regions and networks, in a more global rather than local way.
Here, we will conduct a systematic review of available fMRI data on self-related processing in SZ with respect to the whole brain.The first aim is to identify regions throughout the whole brain that are associated with abnormalities in self-related processing in SZ compared to healthy controls (HC).This aim will enable us to elucidate the exact neural correlates of basic self-disturbance which remain yet unclear.The second aim was to determine how these areas differ based on the task contrast (self-only, self-other, self-baseline) and psychological paradigm (self-reflection, self-referential or self-agency).This is to ensure that the findings are specific to self-disturbance and do not reflect an artifact of the task at hand.
The final aim was to map these areas onto the three-layer topography model of the self (interoceptive, exteroceptive, mental) (Qin et al., 2020) alongside hyperactivation and hypoactivation data (with respect to healthy controls).The recently established three-layer topography model of the self, described by Qin et al. (2020) allows for the psychological characterization of neuroimaging data.The three-layer topography model divides the self into three layers, each one building on the other to conceptualize how the self interacts with the external world.At its core, the interoceptive layer is concerned with internal sensory signals that are a prerequisite for the integration of internal and external stimuli.The exteroceptive layer represents an external stimulus related to one's own body, and it is an intermediate level that connects internal   and external processes.Lastly, the mental layer processes non-bodily signals that are self-related such as personal traits and mental content (Qin et al., 2020).
The thee-layer topography model provides an opportunity to link neuroimaging data to psychological constructs.This can be of great utility in both diagnostic and therapeutic applications in SZ.Therefore, we hypothesized that 1. Various regions throughout the whole brain show abnormalities related to the self, indicating a global rather than local disturbance, 2. That this holds across different tasks related to the self, indicating its specificity to self-disturbance rather than task type, and 3.That all regions of the three-layer topography of self are affected by the changes in SZ, falling in line with a fundamental self-disturbance.

Preliminary search
In a preliminary search (May 2023-September 2023), a small sample of papers were manually selected based on title using the keywords (SELF) AND (fMRI) AND (Schizophrenia) in PubMed.This stage was intended to scope the field and develop the exclusion and inclusion criteria.12 papers were selected in this stage.1 was excluded due to the absence of fMRI imaging, 4 were excluded due to insufficient focus on the self, 2 were excluded due to insufficient healthy control (HC) comparison group and 1 was excluded due to a focus on high-risk rather than clinical patient population.4 papers were included and informed the development of the inclusion and exclusion criteria for further screening (Table 1).Therefore, they were subsequently investigated in depth and included in the data processing phases described in Sections 2.4 & 2.5.

Review search
In a second step, literature reviews (October 2023) were collected to further understand the field and the link between neuroimaging and selfrelated processing.The following terms were inputted into PubMed: (SELF) AND (fMRI) AND (Schizophrenia) AND (Review).42 papers were received, of which only 7 were selected, and an additional paper was manually selected outside of the search.Of those 8, 2 were excluded due to being in French.6 papers were read fully, and the references were carefully considered as potential papers for inclusionsubsequently, 50 papers were selected in this step.After careful title and abstract screening with Table 1 criteria, none of the papers were selected for further analysis.

Systematic search
In a third step, the keywords "(Self) AND (fMRI) and (Schizophrenia)" were entered into PubMed with 447 results.Papers were selected based on the evaluation of title and abstract; 139 were selected.These papers were fully read to ensure they met the inclusion and exclusion criteria (Table 1) established in Section 2.1.Emphasis was placed on the task method to ensure it included self-related stimuli.Any papers that did not include primary data (e.g.reviews) were excluded.The inclusion/exclusion criteria were modified throughout the data acquisition process.The final criteria are presented in Table 1.The screening process can be summarized in the PRISMA figure (Fig. 1).23 papers were selected for data extraction (Section 2.4).scales included information about the scales used for positive and negative symptom characterization and scales for other measures where applicable.Hypotheses of all studies were included.Tasks were categorized into several measures including the differentiation of self-other, self-baseline or self-only contrasts.Other measures included stimuli type (language/text, visual, auditory, or motor), and content (positive, negative, etc.).Measured outcomes were categorized as "Does the stimulus describe (you, other, baseline)", "Is the stimulus (me, other, baseline)", a numerical scale, or a perception about the stimulus itself.Then, each task was categorized as either self-agency, self-referential or self-reflection focused.Self-reflection is related to aspects of self that are more intrinsic.For example, do the following personality traits describe you (Fuentes-Claramonte et al., 2020)?This requires self-reflection on one's characteristics.

Data extraction
Self-referential is more exteroceptive, related to distinguishing stimuli that are relevant to oneself or another entity (Northoff et al., 2006).For example, what is the level of relevance on a scale of 1-3 between this word and a picture of your face (Lee et al., 2016)?Agency is the experience of the self rather than another person causing a certain action or movement (Northoff, 2021).For example, in an experiment, you are asked to keep the cursor moving in a central square on a screen and determine whether the movements are influenced by you or by others (Spaniel et al., 2016) (Refer to Supplementary Data under the "Task" tab for more information relating to tasks of included studies).Neuroimaging data was categorized based on the contrast as ( 1) self (self without specified contrast), (2) self-other (self vs. other contrast) or (3) self-baseline (self vs. baseline contrast).An example of these contrasts would be (1) Does this attribute describe you? (2) Does this attribute describe you or the president (other), (3) Does this attribute describe you vs. how many letters are in the word "red"?(baseline).
These categorizations were inspired by van der Meer et al. ( 2010) who incorporated similar distinctions in their review.However, many categories identified were excluded from the analysis as they were not sufficiently self-focused.These include other-self, other-only, otherbaseline, non-specific and uncategorized contrasts.Therefore, the literature that focused exclusively on such contrasts was excluded from further analysis (Allen et al., 2007;Shergill et al., 2000;Simons et al., 2010;Uhlmann et al., 2021;Vinogradov et al., 2008) (Refer to Supplementary Data for more information).Each coordinate or brain region was selected based on a differential activation between patients and HC.This information was included in neuroimaging data through "Patients >/< HC" labels, which indicate hyperactivation or hypoactivation as present in patients compared to HC.The anatomical term and/or Brodmann Area (BA) were included alongside the coordinates (in MNI or TAL) with a significance level.Additional stimulus information that pertained mostly to valence, content, or additional information was included.

Data analysis
In a data analysis step, all neuroimaging data was separated into their corresponding contrasts (Step 3 in Fig. 2).All coordinates were translated from TAL to MNI to allow for consistency and matched to a corresponding Brodmann Area.Both actions were carried out by an MNI-TAL translator software (The MNI Talairach Tool [WWW Document], n.d.).If the BA was not found using the MNI-TAL tool, then the BA was determined by the identification of a BA in its respective study or based on the anatomical term provided.If such efforts failed to provide an appropriate BA, these coordinates were excluded from the analysis.Then all three contrasts (self, self-other, and self-baseline) were compared to identify areas of overlap.BA's were determined as "overlapping" if they belonged to the 2 or more of the contrasts.If a BA overlapped in 2 or more of the contrasts, they were included for further analysis (Indicated as a red overlay in the Venn diagram in Fig. 2).The BA's that are common among two or more "self" contrasts were then matched to the paradigm type (self-reflection, self-referential, and self-agency) and the three-layer topography model developed by Qin et al. (2020).In the last step, hyperactivation and hypoactivation data were overlayed with respect to each self-layer to determine any patterns of activation that differ between SZ and HC.

Overview of participants in included studies
A total of 23 papers were included for analysis.The demographics of the participants in all studies chosen for inclusion are presented in Fig. 3. Patients, who had a diagnosis of schizophrenia as determined by inclusion criteria, numbered a total of 471 individuals; 308 of which were male and 143 of which were female.The patients were 35.63 years old on average and medicated.The healthy controls (HC), all of which have an absence of any psychiatric diagnoses, presented with similar demographic information.HC totalled 431 individuals, 251 of which were male and 169 of which were female.Additionally, HC were 34.31 years of age on average.Hence, no major differences between healthy controls and schizophrenia patients were appreciated (Fig. 3).

Overview of task design in included studies
Fig. 4 outlines the different tasks performed in each study.Most studies (14/23) incorporated self-other and self-baseline contrasts in their task.This is a beneficial characteristic of a task design, considering it allows for multiple dimensions of the self to be studied, beyond selfother differentiation.As for the stimuli used in the tasks, the majority were language/text stimuli (16/23) which involve the presentation of a word or phrase, commonly a trait adjective.Other stimuli include auditory (5/23) and motor/body ownership (2/23).The variety of different stimulus types can allow for the study of self-related processing more comprehensively.The measure refers to the task output or participant response regarding the stimuli.These showed slight variations between studies which may influence neuroimaging data through the recruitment of differential brain networks.Lastly, the self-focus type relates to the nature of the task and whether it is self-reflection, selfreferential or self-agency.This step is used for further analysis in Section 3.4 to determine whether abnormal neuronal recruitment in SZ is consistent across self-related tasks.

Identification of BA's involved in self-related processing
In Table 2, the BA's that exhibit differential activation between SZ patients and HC were categorized based on involvement in self-other, self-baseline, and self-only differentiation.It is evident that the areas showing differential activation between SZ and HC, in all three contrasts, span the whole brain and include regions from the frontal, parietal, temporal, occipital and limbic lobes.To identify regions for further analysis, the BA areas were all inputted into a Venn diagram (Fig. 5A), and regions that corresponded to at least two or more of the contrasts were selected for further analysis.Some regions overlap fully on three contrasts (BA 6,9,23,40) whereas others only correspond to two contrasts (BA 4,7,8,10,31,39) (Fig. 5B).The overlap in BA's across different self contrasts (e.g.self-other, self-baseline) provides more support for the role of these regions in self-related processing.By extension, this supports the notion that these regions involved in self-related processing are disrupted in SZ as compared to HC.

Mapping BA's to different psychological paradigms
The BA's that overlapped in two or more contrasts were then mapped to three psychological paradigms identified.(Fig. 4) BA's 10 and 23 showed overlap in all paradigms indicating a central role in self-related processing.BA 8,39 show overlap in self-reflection and self-referential paradigms, BA 11,31 show overlap among self-referential and selfagency, and BA 40 show overlap in self-agency and self-reflection (Fig. 6 Bi).When mapping the BA's specific to the psychological paradigms, more areas are abnormal in self-reflection related tasks in comparison to self-referential and self-agency tasks.This may reflect a larger amount of studies employing self-reflection tasks as compared to the others (Fig. 4); however, the results still imply a core set of BA's, specifically those that are presented in Fig. 6 Bi, that have consistently presented as abnormal both with respect to stimuli-differentiation and paradigm task.

Mapping BA's to the three layer topography model of the self alongside hyperactivation and hypoactivation data
In the context of the three layer model of the self, BA's 4,11,18,20 and 23 did not correspond to any layers as per Table 3 in Qin et al. (2020).The BA's that correspond to a layer of the three-layer model mostly correspond to the mental and exteroceptive self, and less so to the interoceptive self.This may indicate more neural abnormalities in the regions associated with the mental and exteroceptive self in SZ as compared to HC. BA 40 was identified as abnormal in the exteroceptive and interoceptive self, and BA 10 was identified as abnormal in both the mental and exteroceptive layers.Therefore, through incorporating three filters on data presented in Table 2, BA 10 and 40 have been consistent in exhibiting abnormalities in activation in SZ compared to HC.Additionally, Fig. 7 indicates that self-disruption in SZ may arise from abnormalities primarily in mental and exteroceptive processing.
To extend these findings further, the abnormalities can be characterized as hyperactive or hypoactive in SZ with respect to HC. BA 10, 8, and 6 show hyperactivation in the mental self, BA 10, 7, and 40 show hyperactivation in the exteroceptive self alongside hypoactivation in BA 32, and lastly, BA 40 shows hyperactivation in the interoceptive self.In particular, BA 10 and BA 40 exhibit hyperactivation in SZ as compared to HC.

Identification of BA's involved in self-related processing
To identify areas of relevance for self-related processing, BA's belonging to a minimum of two contrasts (self-only, self-other and selfbaseline) were selected for further study.The BA's that satisfy those requirements span a diversity of functions including motor (BA 4,6),visual (BA 18,20), language (BA 7,8,39,40), executive (BA 9,10) and emotion (BA 23,31,32) (Guo et al., 2022).Although BA 11 is indicated as a region of overlap between two contrasts, its function is unclear based on the available literature.However, it seems to be involved in olfaction and reward (Ernst et al., 2004;Miyanari et al., 2007;Rogers et al., 1999;Royet et al., 1999).The BA's that belonged to one contrast only were excluded from further analysis.These areas include the insula (BA 13), visual (BA 17,37), emotion (BA 24), and language (BA 21,22) regions (Guo et al., 2022) which may represent non-self-related Fig. 4. Task summary schematic.The Y-axis describes the four characteristics by which tasks were categorized, by contrast (both, self-other, self-baseline), stimulus type (language/text, auditory, motor/body ownership), measure, and self-focus type (self-reflection, referential, agency).The X-axis includes all 23 papers included for analysis.Created with BioRender.com.

Table 2
Summary of neuroimaging results as categorized by contrast (self-other, self-baseline, self-only).Each coordinate is numbered with information on the hemisphere (R for Right, L for Left) and the label associated with the Brodmann Area (BA) as determined using the MNI-TAL conversion tool (The MNI Talairach Tool [WWW Document], n.d.).All coordinates are presented in MNI, and the study from which each coordinate was extracted is provided in the rightmost column.(Jardri et al., 2011) thinking, an artifact of the task employed, or a lack of representation in the available literature.Nevertheless, these functions are still largely represented in the areas included for further analysis.Here, we show that regions disrupted during self-related processing in SZ are generally distributed across the entire brain (Fig. 5B).This includes higher-order transmodal regions in the cortical midline structures (Potvin et al., 2019;Shad et al., 2011;van der Meer et al., 2010) as well as lower-order regions like visual (BA 18,20) and motor regions (BA 4,6).This suggests a global neural correlate of the self-disturbance in SZ encompassing the whole brain rather than specific local regions or networks.These global spatial changes are well in accordance with analogous global temporal changes during self-tasks in SZ (Hamilton and Northoff, 2021).Hence, the self-disturbance in SZ seems to affect the whole brain in a global way in both temporal and spatial regards.

Mapping BA's to different psychological paradigms
When categorized based on psychological paradigm type (selfreflection, self-referential, and self-agency), BA 10, 23 showed differential activation between SZ and HC in all paradigms (Fig. 6A).BA 10, a component of the frontal pole, activates during cognitively intensive tasks (Snelleksz et al., 2022).BA 23, or the posterior cingulate gyrus, is likely involved in functions that rely on spatial memory and visuospatial orientation, like the navigation of the body in space (Holt et al., 2011).Holt et al. hypothesized that SZ patients abnormally recruit the posterior cingulate (BA 23) to rely on visuospatial and episodic memories to judge the relevance of self-descriptive adjectives.Whereas HC recruit frontal regions, like BA 10, to retrieve and store semantic aspects of the self.This may suggest greater reliance on external stimuli in SZ patients when making judgments about self-stimuli.
Other areas of importance include language (BA 8,39,40) and emotion (BA 31) related areas (Guo et al., 2022) that may indicate a dysfunction in language and emotion processing that is fundamental across different psychological paradigms.This is a promising finding that supports the existence of self-dysfunction in SZ that is independent of psychological paradigm types.Taken together, the abnormal recruitment of language and emotion areas can contribute to selfdisturbance alongside BA 10 and 23 to precipitate psychopathological symptoms like delusions, hallucinations, and misattributions (Menon et al., 2023).
However, there are many regions unique to self-reflection (Fig. 6A), earlier proposed as a potential consequence of the larger number of studies utilizing the paradigm.The regions unique to self-reflection include motor (BA 4,6), language (BA 7), executive (BA 9), and visual (BA 18,20) functions (Guo et al., 2022).Other potential reasons for this finding are that self-reflection tasks may be more general or non-specific enabling them to recruit non-self-related regions.Additionally, selfreflection tasks may be difficult to differentiate from self-referential tasks due to subtle differences, causing an overrepresentation of studies included as self-reflection.This is an inherent limitation of the present study and presents a need to characterize psychological paradigms more concretely.This would enable task-based imaging to be compared more systematically.However, one promising observation is that the self-reflection tasks seem to employ similar measures and stimulus types (Fig. 4).This indicates that the regions unique to selfreflection are likely not attributable to diversity in task type but rather to the paradigm as a whole.

Table 3
Summary of neuroimaging results as categorized by the three-layer topography model of self from Qin et al., 2020(Qin et al., 2020) (interoceptive, exteroceptive, mental).The BA's were noted alongside their corresponding coordinate numbers from Table 2. Coordinates are separated based on whether they were hyperactive or hypoactive in SZ patients as compared to HC.Some BA's repeat due to their association with multiple paradigms as indicated in the rightmost column.BA's that did not correspond to any layers of the three-layer topography model are indicated under the heading "no correspondence."We can extend these findings by overlaying the neuroimaging data to the three-layer topography model of selfwhich is categorized into the three levels of the self (interoceptive, exteroceptive, and mental).Regions showing differential activation between SZ and HC are present in all three layers of the self (Table 3 & Fig. 7).These include BA's 7,9, and 32 corresponding to the exteroceptive layer, BA's 6,8,31, and 39 corresponding to the mental layer, and two areas that belong to two layers -BA 10 (mental and exteroceptive) and BA 40 (exteroceptive and interoceptive).When moving from the interoceptive to the mental layer, the areas implicated in differential activation between SZ and HC span more functions.This is in line with the notion that these layers build on each other from interoceptive to mental with each layer encompassing more regions.Indeed, the mental layer contains regions of differential activation between SZ and HC corresponding to the most functions including motor (BA 6), language (BA 8,39), executive (BA 10) and emotion (BA 31) (Guo et al., 2022).The exteroceptive layer falls next with language (BA 7,40), executive (BA 9,10) and emotion (BA 32).Then followed by the interoceptive layer that includes only BA 40 corresponding to language function.Within the context of hyper/hypoactivation, the exteroceptive and interoceptive layers primarily indicate hyperactivity in SZ as compared to HC (Fig. 8).This finding is consistent with our previous discussion of over-reliance on external stimuli during self-processing (via BA 23) in combination with dysfunctional stimulus integration (via mental processing) that can manifest as SZ symptomology -particularly hallucinations, delusions and misattributions.This finding can support the utility of the three-layer topography model in linking neuroimaging data with the psychopathology of SZ.
Together, our data show the global nature of the self disturbance in SZ affecting various regions throughout the whole brain holding across different tasks and layers of self.Albeit tentatively, we suppose that such global neural disturbance of self is related to what has been described as basic self disturbance in phenomenology (Hamilton and Northoff, 2021;Henriksen and Parnas, 2012;Martin et al., 2023;Sandsten et al., 2022;Sass and Parnas, 2003).In the same way the experience of an altered sense of self affects and reverberates more or less through all symptoms, the brain's global changes during self tasks seem to affect all regions including higher-order transmodal and lower-order unimodal regions.This is further amplified by the observation of the psychological unspecificity of these global changes with respect to different cognitive self-tasks and self-layers.The basic disturbance of self, as described in phenomenology, may thus be mediated by the brain's global disturbance in both its topography and dynamics (Hamilton and Northoff, 2021), whose underlying basic neural disturbance remains yet to be identified.Given that both topography and dynamics reflect the brain's spatial and temporal organization, one would tentatively suppose such basic neural disturbance to be primarily spatiotemporal which should then be manifest in corresponding changes in time and space experience and perception of SZ (Arantes-Gonçalves et al., 2022;Hirjak et al., 2013;Stanghellini et al., 2020;Stanghellini et al., 2016).These entails have recently been introduced as 'Spatiotemporal psychopathology' (Hirjak et al., 2023;Northoff, 2016aNorthoff, , 2016b;;Northoff et al., 2023;Northoff and Hirjak, 2022).

Conclusion
Schizophrenia is characterized by a basic self-disturbance whose neural correlates remain yet unclear.Conducting a systematic fMRI review, we show abnormalities in a variety of both higher-order transmodal and lower-order unimodal regions during self-tasks in SZ.This holds across different self tasks like self-reflection, self-referentiality and self-agency.Moreover, these neural changes are observed across different facets or layers of self with regions predominantly related to the mental-self and exteroceptive-self exhibiting hyperactivity.We therefore assume a more global whole brain rather than local regional correlate of the self-disturbance in SZ which, albeit tentatively, converges with the 'basic self-disturbance' as described in phenomenology.That opens the door for an integrated brain experience symptom approach as suggested by 'Spatiotemporal psychopathology' (Hirjak et al., 2023;Northoff, 2016aNorthoff, , 2016b;;Northoff et al., 2023;Northoff and Hirjak, 2023).The utility of spatiotemporal psychopathology is in understanding the "common currency" of the brain's neural activity and the psychopathological symptoms (Northoff, 2022).Deciphering this common currency brings us one step closer to the implementation of neuroimaging for the diagnostic and therapeutic advancement of SZ and psychiatric disease as a whole.

Declaration of competing interest
There are no conflicts of interest to report.G.N. is supported by Canadian Institutes of Health Research, Social Sciences and Humanities Research Council, Natural Sciences and Engineering Research Council, New Frontiers in Research Fund, and the UK-Canada AI grant which he holds together with Karl Friston.The funding sources had no involvement in neither the collection, analysis and interpretation of data, nor in the writing of the article nor in the decision to submit the article for publication.

Fig. 1 .
Fig. 1.PRISMA flow chart for search strategy and literature screening.
S.G.Sabbah and G. Northoff Data extraction was performed using Microsoft Excel, (See Supplementary Data) with 5 tabs titled: Demographics, Psychological Scales, Hypothesis, Task, and Neuroimaging.The demographics tab included information on study identification, Patients (n), Type and Diagnosis of Schizophrenia, Exclusion Criteria, Age Mean, Sex (M/F), Medications and HC [n, Age Mean, Sex (M/F), Exclusion Criteria].Psychological

Fig. 2 .
Fig. 2. Schematic for data analysis of brain region coordinates collected from 23 studies.Coordinates refer to brain regions that exhibit differential activation in schizophrenic patients as compared to HC. Created with BioRender.com.

Fig. 3 .
Fig. 3. Aggregated data on the demographics for patients and healthy controls (HC) in the 23 papers collected for inclusion.Dx = diagnosis.Created with BioR ender.com.

Fig. 5 .
Fig. 5. A) Venn Diagram of Brodmann areas that present abnormalities in selfother, self-baseline or self-only.Areas that overlap in 2 or more contrasts are selected for further investigation.B) BA regions that overlap in two or more contrasts mapped onto a lateral and medial view of a brain sectioned off into different BA's. Figure adapted by author from Huyck and Passmore (Huyck and Passmore, 2013).Created with Biorender.com.

Fig. 7 .
Fig. 7. A) Venn Diagram of Brodmann areas that present abnormalities among the three-layer topography model of the self (interoceptive, exteroceptive, mental).Refer to Fig. 2 for a schematic of the three-layer topography of self model.B) BA regions that overlap in two or more layers as well as those present only in one layer mapped onto a lateral and medial view of a brain sectioned off into different BA's. Figure adapted by author from Huyck and Passmore (Huyck and Passmore, 2013).Created with Biorender.com.

4. 3 .
Mapping BA's to the three layer topography model of the self alongside hyperactivation and hypoactivation data

Fig. 8 .
Fig. 8. BA's associated with the three-layer topography model specific to each layer mapped onto a lateral and medial view of a brain sectioned off into different BA's.Green highlighted regions indicate hyperactivation in SZ patients with respect to HC and vice versa for red.Brown highlights indicate equal hyperactivation and hypoactivation in SZ patients with respect to HC. Figure adapted by author from Huyck and Passmore (Huyck and Passmore, 2013).Created with Biorender.com.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 1
Exclusion and inclusion criteria for screening of literature.