First-Rank Symptoms in Schizophrenia: Reexamining Mechanisms of Self-recognition

Abstract

Disturbances of self are a common feature of schizophrenic psychopathology, with patients reporting that their thoughts and actions are controlled by external forces, as shown in first-rank symptoms (FRS). One widely accepted explanatory model of FRS suggests a deficiency in the internal forward model system. Recent studies in the field of cognitive sciences, however, have generated new insights into how complex sensory and motor systems contribute to the sense of self-recognition, and it is becoming clear that the forward model conceptualization does not have unique access to representations about the self. We briefly evaluate the forward model explanation of FRS, reassess the distinction made between the sense of agency and body ownership, and outline recent developments in 4 domains of sensory-motor control that have supplemented our understanding of the processes underlying the sense of self-recognition. The application of these findings to FRS will open up new research directions into the processes underlying these symptoms.

Introduction

Anomalies of self-experiences are a critical feature of schizophrenia-spectrum disorders. First-rank symptoms (FRS), particularly, are considered to be at the core of schizophrenia psychopathology.¹ FRS refer to experiences whereby thoughts and actions are perceived to be under the control or influence of an external agent or where there is a loss of clear boundaries between the sense of self and others. In contemporary research classifications, a subset of FRS are typically examined, frequently referred as “passivity experiences” (see table 1). People experiencing FRS suffer from deficits in self-recognition because they have difficulties distinguishing between self and others and are unsure whether thoughts and actions are separate and independent from external influences. FRS are intriguing because they raise questions about how the human mind is able to make a distinction between self and nonself. Unraveling the processes that underlie these symptoms has relevance beyond the immediate theoretical implications for schizophrenia by shedding some light on the human abilities that form the basis of all communication and socialization.

Table 1.

First-Rank Symptoms (Passivity Symptoms) in Schizophrenia

Symptom	Definitiona	Patients’ Descriptionb
Third-person auditory verbal hallucinations	Voices which have no real origin in the outside world and which the patients regard as separate from their own mental processes. Voices speak about or comment on the patient.	I hear the voice of my stepfather saying “he never gets anything right.” The voice of God says to me “she is the only one who can save us now.”
Loud (audible) thoughts	Patients report that their own thoughts seem so loud that someone near by could hear them.	“I can hear myself think. Loud enough for another to hear.”
Thought insertion	Thoughts are experienced as not being the patients’ own. Often, the alien thoughts are perceived to have been inserted from outside by an external agent.	“Some thoughts are not my own. They belong to Satan: He types in evil thoughts about killing people.” “People are stuffing around in my mind, putting thoughts inside.”
Thought broadcast	Patients experience their thoughts as accessible to others, sometimes diffusing out of their mind, so that the content of the patient's mind is no longer private.	“Other people can know my thoughts. I am not able to feel any privacy.”
Thought withdrawal	Patients report that their thoughts have been taken out of their minds. The thoughts are experienced to be actively extracted by others.	“Satan rips out, pulls out, brings out the thoughts out of my soul. It feels like that.” “My thoughts get stolen”.
Made experiences/ replaced control of will	Actions, intentions, and/or feelings are experienced to be under the control of some other force. Often, patients believe that someone else's words are coming out using their voice or that their movements are being willed by some outside power and that their personal intention has been replaced by the sense that some other is in control.	“Someone makes me move my mouth, my jaw, and pushes the words out. I can't help but say something.” “I'm forced to walk around, I'm being made to turn right and left.” “I don't have control over my feelings any more, how I react to emotions now is not the same as before I entered the house of ghosts”.

^aBased on the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) definitions.⁷²

^bWaters et al.⁶

FRS play a key role in the diagnosis of schizophrenia and feature prominently in current diagnostic systems. The growing literature on FRS reflects the renewed interest in their cognitive and neural processes^2,3 and their diagnostic significance.^4,5 FRS, however, are not easily explained in terms of traditional neuropsychological dysfunction,⁶ and the pathophysiological mechanisms underlying these symptoms continue to elude researchers. One widely accepted explanatory model of FRS suggests a dysfunction in the internal forward model system.^7,8 The model suggests that the ability to differentiate between the self and others is rooted in motor control processes and that FRS arise because of faulty action processing mechanisms. This model has produced interesting findings but has tended to dominate experimental research into FRS. There has been a paucity of empirical investigations questioning this view, despite accumulating evidence that the forward model does not provide a definite answer to FRS. In addition, recent studies in the field of cognitive sciences have generated new insights into how complex sensory and motor systems contribute to the sense of self-recognition. Studies reveal, firstly, that there are many different, overlapping, representations of the self that are not exclusively tied to the motor system and, secondly, that there is a range of cues available to recognize the self that have yet to be individually examined in patients with FRS. For the purpose of this overview, we adopt a broad definition of self-recognition, namely the recognition of the self as an entity independent from others and as an agent of behavior, given that concepts and terms referring to this area of inquiry often overlap.

In the current article, we briefly evaluate the forward model explanation of FRS and review recent advances in 4 areas of sensory-motor control that have supplemented our understanding of the mechanisms of self-recognition in healthy people. These new research directions may guide future investigations into the processes that underlie FRS.

Forward Model Framework

The forward model is a theoretical model of cognitive motor control, whose role consists in predicting the sensory consequences of motor commands. This predictive process allows the differentiation of sensations occurring as a result of intentional movements from sensations originating from changes in the external world.^9–12 The ability to distinguish between these sensations arises because the forward model anticipates, and diminishes, the sensory effects occurring as a result of self-generated movements. The internal system deduces from diminished sensory effects that sensations have arisen from a motor command that we issued. By contrast, sensations that are externally produced are experienced more strongly and are interpreted to have an external cause. Frith's proposal^7,8 is that patients with FRS are unable to predict internally the perceptual consequences of their actions so that the salience of sensations is not diminished, resulting in actions being perceived to be externally generated. Forward model abnormalities are also thought to predict difficulties in self-monitoring of willed intentions and actions, and deficits in detecting and correcting errors during self-generated movements. Jeannerod^13,14 has recently expanded on the forward model by proposing an action simulation framework. The proposal not only retains the idea that the motor system uses internal forward models, but it also suggests that internal systems compute different kinds of representations for first-person and third-person information that enable self-/other distinctions and external attributions.

One difficulty facing FRS research is their heterogeneity in presentation, which includes abnormalities of thoughts, action, perception, and emotions. To account for this diversity, the forward model makes the 2 further assumptions: (1) that FRS represent one core dimension related to cognitive motor control, resting on the idea that thoughts and actions are all instances of intentional behaviors ranging from the covert (mental activities) to the overt (actions)¹³ and (ii) that these phenomena and behaviors are subject to the same constraints as motor processes and consequently that the outcome of a disorder in motor control, notably in forward modeling, comprises anomalies spanning the entire range of human experiences.

In summary, the forward model suggests that the sense of self is underpinned by a complex multistage motor control system. A dysfunction in this system is thought to be responsible for creating the self-disturbances shown in patients with FRS.

Evidence of Forward Model Dysfunction in FRS

Evidence cited in support for the forward model framework includes studies showing that schizophrenia patients with FRS demonstrate abnormalities in sensations arising from intentional movements, as shown on a task measuring responses to self-produced tactile sensations (tickles)¹⁵ and on a task assessing the cancellation of retinal information during smooth eye pursuit.¹⁶ Studies have also demonstrated abnormalities in self-monitoring mechanisms, such as failing to correct a wrong movement in the absence of visual feedback.^17,18 Patients with FRS also show difficulties when making decisions based on whether the visual feedback of an action corresponds to their own movements, or to another person, as demonstrated on self-/other attribution tasks,^19,20 and in remembering the origin of verbal material in memory.^21,22

Several limitations, however, are impeding an evaluation of the strength of the association between a forward model dysfunction and FRS. For instance, many studies have reported forward model impairments in the broader schizophrenia population without reference to FRS, so that the association of these deficits to FRS specifically remains under question. Other methodological limitations in the literature include (1) inconsistency, or lack of transparency, concerning which symptoms are targeted during the patient selection process, (2) small sample sizes, and (3) inclusion of nonpassivity phenomena (such as delusion of mind reading) in FRS categorization. Another limitation is the use of a recruitment criterion based on the presence of only one FRS. Given the heterogeneous nature of FRS, and in view of the risks of incorrectly identifying these symptoms, studies may reach incorrect conclusions by indexing symptoms other than FRS. A recruitment criterion based on the presence of at least 2 FRS might provide a more reliable categorization method. A suitable comparison group may include patients with no lifetime history of FRS.

Theoretical aspects of the model have also attracted criticism.^23–26 For example, questions have been raised concerning the correct interpretation about the consequence of a forward model dysfunction, given that forward model deficits may account for the feeling of not being in control of one's actions but not for the experience that a separate agent is controlling these actions.²⁴ In addition, Blakemore et al²⁷ have showed that the feeling of not being in control is not necessary to generate the belief that one's actions are controlled by external forces. Also, a complete failure of the forward model would result in broad deficits in motor control, while studies have shown that forward model processes for low-level automatic action control seem to be unimpaired in schizophrenia.²⁸

Few studies of FRS, however, have considered that the forward model conceptualization is only one of several sources of information that contribute to self-recognition. One possible reason comes from the prevalent view that patients with FRS have a disordered sense of agency but an intact sense of body ownership. This has resulted in an almost unilateral research focus on motor control processes in FRS.

Sense of Agency and Body Ownership

The observation that patients with FRS do not have the sense of causing their movements and frequently misattribute their own actions to another agent has led to the proposal that FRS reflect a disturbance in the sense of agency.^8,24,29,30 By contrast, patients with FRS appear to have a preserved sense of their own body, resulting in the proposal that they have an intact sense of body ownership. The terms have been given different definitions, but the sense of agency may be defined as the sense that one is in control of one's actions or “the sense an agent has that he or she is the author of that action.”³¹ This differs from the sense of body ownership, which refers to a sensory awareness of the body that is not exclusively tied to actions. Tsakiris et al³² have defined body ownership as “the special perceptual status of one's own body, which makes bodily sensations seem unique to oneself.” In this conceptualization, the sense of agency is intrinsically associated with, and secondary to, action execution and therefore linked to the motor control system. By contrast, body ownership is tied to the combination of sensory experiences arising from the body senses, as well as internal information about the body's position at all times.

The theoretical differences between the sense of agency and body ownership have attracted much interest,³³ but several lines of evidence are now challenging the view that patients with FRS have a selective deficit in the sense of agency. To start with, the sense of body ownership and the sense of agency are not modular and discrete but may, in fact, be interdependent and overlapping.^31,34–36 For example, the computation of action processes contributing to a sense of agency relies critically on body ownership processes, such as coherent multisensory signals arising from the same body parts, as well as internal spatial body representations. The body senses and position of the limbs are also constantly being updated during the body's interaction with the environment, and this continuously changing state of reference is used by the forward model to recalibrate its predictions about the state of the system and to differentiate between self- and other-generated sensations,^26,33,37 demonstrating the interdependence of body ownership and agency processes.

Deficits on attribution tasks shown in patients with FRS, in which patients are asked to identify a movement as belonging to the self or to another person based on video visual feedback,^19,20 have also been cited as evidence of a selective impairment in the sense of agency. However, performance on these tasks necessarily requires an implicit knowledge of one's body as a sensory object that is moving (ie, a sense of body ownership), so that attribution deficits may also reflect a disturbance in body ownership.³⁸ In addition, low-level perceptual deficits in combining visual and somatosensory (ie, nonmotor) signals from the same body parts may also affect the ability to judge accurately whether an action has been performed by the self or by another person, so that an explanation based on sensory deficits may also account for attribution errors. Conversely, the ability to make reference to the sense of body ownership actually involves making a self-attribution judgment,³⁸ which therefore does not only arise from a sense of agency. Altogether, this suggests that the concepts of agency and body ownership are intrinsically linked and cannot be fully dissociated.

Philosophical discussions have also been helpful in formulating aspects of self-experiences. For instance, phenomenological descriptions of the experiential self have proposed levels of selfhood that are hierarchically organized and overlapping.^39–41 Briefly, there are different forms of selves, starting from a very basic notion of the self that is prereflective and determined by an implicit and sensory awareness of first-person experience. This form of self-awareness may be linked to the body ownership experience. By contrast, the sense of agency may be tied to a more explicit and reflective form of the self that is consciously defined as “I.” Finally, there is a more complex level of selfhood that is composed of social and personal characteristics. Importantly, these various levels of selves are entwined and cannot be observed in isolation from each other.

Furthermore, a dissociation between the sense of agency and body ownership has yet to be tested unequivocally in schizophrenia patients with FRS. Instead, phenomenological examinations have pointed out that patients with schizophrenia suffer from disturbances in the sense of self that extend beyond the domain of action awareness, comprising the entire field of self-experiences including abnormalities in the sense of body ownership.^23,42 In sum, there is a need for experimental studies to challenge the views that FRS arise exclusively because of a dysfunction in motor control processes, and that they are essentially a disorder of agency.

Cues Underlying Self-recognition

A different approach to understanding the processes that contribute to self-recognition has consisted in using healthy people to examine how the self is inferred, utilizing experimental paradigms in which self-distortions are experimentally manipulated. The cues examined refer to body ownership processes (sensory integration and body schema), motor control subcomponents (efferent motor signals), and subjective psychological experiences. While the forward model includes many of these signals as constituent processes in its proposal, the critical point is that the studies reviewed below demonstrate that these cues contribute in their own unique way to the sense of self-recognition. This means that the sense of self is neither exclusively linked to a combination of processes tied to the motor system nor is it secondary to action execution, so that deficits in nonmotor processes can also result in abnormalities of self-experiences. Consequently, the forward model formulation, as it is conceptualized, does not have sole access to representations about the self.

Importance of Sensory Integration

While the forward model emphasizes the role of motor feedback and programming, new research is showing that processes extending beyond intentional motor processes are also important contributors for a sense of self.^13,43–46 For instance, recent studies have highlighted the importance of the successful integration of multiple sensory signals arising from the external environment. Matching visual and somatosensory input from the same body parts are critical for building a coherent image of our body, as shown in experimental paradigms where inducing a sensory conflict is sufficient to produce a dissociation in the sense of self. In one such study, a set of rubber hands is placed in a position that is compatible with the participant's posture, while the participants’ hands are out of view. Watching brush strokes applied to the rubber hands while feeling synchronous strokes on one's arm induces the feeling that the rubber hand is part of the participant's own hand.⁴⁴ These results suggest that altered self-experiences can arise as a result of conflicting/incongruent sensory signals. In the same way, out-of-body experiences may be created experimentally simply by manipulating the senses.^45,47 Ehrsson⁴⁵ recently conducted an experiment in which people watched an illusory image of themselves through virtual-reality goggles that projected an image of themselves from the back. When the experimenter stroked the person's back at the same time as projecting the image of a plastic rod stroking the illusory body, participants reported the feeling of watching themselves from the back. That is, their sense of self had “transported” itself outside their physical body, so that participants experienced watching another person's body in one's own body. These experiments demonstrate a critical role for multisensory integration in assisting perceptual decisions about self-recognition and show that a mismatch between visual perspective and proprioceptive signals is enough to induce disturbances in self-experiences. Importantly, there was no intentional motor component involved in the paradigm, suggesting that intentions do not have privileged access to representations of self, contrary to the assumptions of the forward model.

Role of Body Schema

The body schema is a dynamic spatial coding of the body, its internal organization, boundaries, and relationships.⁴⁸ It is an internal frame of reference about the structure and position of the body, and it is thought to be important for enabling self-recognition. The body schema is also used to coordinate actions and to allow fluid movement through space by providing a representation of the body parts at all times with respect to each other.⁴⁹ Costantini and Haggard⁵⁰ demonstrated the importance of internal limb representations arising from the body schema using the rubber hand illusion. Specifically, spatial mismatches between the participant's hand posture and the rubber hand posture were created by rearranging either the subject's hands or the rubber hands to be at odds with each other. It was revealed that the body illusion was strongest when the image of the rubber hands matched the participant's existing representation of their own body and lowest when the rubber hands’ position did not correspond to the subject's actual hand positions, underlining the importance of preexisting internal body representations. The authors argued that these representations are needed to provide a spatial context in which to assimilate novel sensory information and constrain motor commands. Importantly, by providing this frame of reference about existing internal body representations, the body schema is thought to contribute to the awareness of one's own body and actions and to the discrimination of self from others. The body schema is, therefore, intimately related to self-recognition.

Importance of Efferent Motor Signals

The forward model proposes that the sense of self is secondary to action execution, yet another point of view argues that efferent (motor) commands play a more specific role toward self-recognition.^32,51,52 Tsakiris et al⁵¹ used a self-recognition paradigm in which subjects could see their right hand as an image via video feedback, while the hand was passively moved either by the subject's left hand or by the experimenter. The results showed that self-recognition was at chance level if only proprioception was available (when the hand was moved by the experimenter), whereas the added efferent information (when the participant made his own hand move) significantly increased self-recognition performance. The authors interpreted the results as suggesting that efferent signals provide more critical cues in enabling self-recognition than proprioception alone, arguing against the forward model proposal that a cohesive sense of self only occurs after sensory feedback. The role of efferent signals was further examined in a manipulation of the rubber hand illusion in which both passive and active finger movements were used.³² In the passive condition, the illusion was localized only to the area that was stimulated (the finger). By contrast, when active finger movements were used, the illusion extended to the whole hand suggesting that efferent motor commands integrated different body parts into a wider awareness of the body in action. The authors proposed that efferent motor signals contribute to the sense of self-recognition by enabling a global subjective awareness of all limbs involved in movement. Importantly, as pointed out by Schwabe and Blanke,⁵³ this means that the role of the forward model goes beyond attenuating sensations of the moving body part by affecting a more global representation of the body in a very complex way.

Role of Psychological Binding Effects

A related area of research has shown that there are subjective effects of actions that also contribute to the understanding of how self-recognition is inferred.^54–58 For example, recent studies from Haggard and colleagues^55,56 have shown that when a voluntary action is followed by an expected sensory consequence, a psychological binding effect occurs so that these events are perceived to have occurred closer together in time than they actually did. In other words, this “intentional binding” effect refers to the subjective perception that cause and effects are drawn together in perceived time. Haggard has argued that this temporal shrinkage between intentional actions and their behavioral effects creates an experience of causation between actions and their consequences, thus enhancing our feeling of self-authorship. From the idea that intentions play a critical role in representing our movements, the authors argue that self-recognition is inferred from this enhanced subjective experience of intentionality. Furthermore, this temporal binding effect is thought to be sensitive to top-down influences associated with motor intentions, suggesting that psychological experiences of anticipated actions have distal effects on sensory-motor perception and that expectations may play a critical role in representing our movements.

Summary and Discussion

Proponents of the internal forward model suggest that a disorder in the system that predicts the sensory consequences of one's actions may result in a disordered sense of self, as in FRS. However, recent developments in the area of sensory-motor control have shown that a model based on motor feedback and programming does not fully account for the clinical and experimental data and that there are several additional cues that contribute to self-recognition: there is a need for consistent and coherent signals across the body senses, as well as information about body location and orientation provided by the body schema. The role of purely efferent motor commands has also been readdressed and is now thought to contribute to body representations by providing a broad subjective sense of our limbs in movement. In addition, it is becoming clear that intentional actions produce the subjective experience that willed actions and their external consequences are related, as shown when voluntary movements and their effects are perceived to have shifted toward each other in time.

Sensory integration and body schema are areas of research that remain to be fully examined in schizophrenia patients with FRS, but there are indications that these processes may be impaired in this patient group. For example, a study by Peled⁵⁹ showed that schizophrenia patients felt the rubber hand illusion stronger and faster than normal controls, although the results were not examined in relation to symptom clusters. More recently, Farrer and Franck⁶⁰ used a modified version of an action attribution task that involved rotated visual feedback of hand postures. They reasoned that if patients with FRS have body ownership impairments, they would show increased attribution impairments as the rotation increased. The authors found that both patient groups with and without FRS made more errors across all conditions compared with a nonclinical control group. Although there were no significant differences between patient groups, and no effect of rotation, patients with FRS actually tended to make more errors across all conditions than patients without these symptoms and controls. This points to difficulties in matching the congruence of tactile and visual signals and deficits in body ownership processes in FRS patients. These results are consistent with studies showing body schema impairments in the wider schizophrenia population.⁶¹ In sum, there are indications that multisensory integration and body schema may be impaired in schizophrenia. Future studies should examine this pattern of performance in FRS patients specifically. Purely efferent motor signals also remain to be investigated in this patient group.

There has recently been evidence of an association between impaired intentional binding and schizophrenia, with patients showing excessive binding between intentional actions and their sensory consequences (Haggard et al, 2003). It was suggested that this contracted sense of time might result in an erroneous perception of action causation and in particular to abnormal action attribution in schizophrenia patients with self-disturbances. However, no symptom analysis was conducted, so the association between impaired action timing and FRS still remains to be tested. However, there is evidence of timing impairments extending beyond overt motor control in schizophrenia patients with FRS as shown on a task of motor imagery^62,63 and on an auditory discrimination task requiring judgments of temporal intervals.⁶⁴ Given the role of timing precision for shaping the sensory awareness of movements⁶⁵ and for generating the subjective perception that some events belong together,⁵⁵ these abnormalities in timing awareness in patients with FRS may contribute to a breakdown in the perceived relations between actions and their effects. How this fits exactly in relation to other processes involved in self-recognition is yet to be determined.

In summary, we have demonstrated that there are several sources of information that contribute to self-recognition. They are not evidence against the forward model framework, given that sensory signals, efferent motor information, expectations, and timing are all constituent parts of the model. However, the above studies show that a motor control theory reliant on a combination of these cues overlooks their unique contribution for enabling self-representations. We have also argued against the prevailing view that FRS are evidence of a selective deficit in the sense of agency because evidence suggests that the sense of agency and body ownership are intrinsically linked and cannot be observed in isolation from each other. Given the lack of clear separation between different forms of selves, FRS should perhaps be viewed as reflecting an altered sense of self-experience (rather than impaired agency).

One last comment about the above research is that it has focused almost exclusively on self-recognition of body experiences and actions. There are concerns²⁴ that other forms of intentional experiences, especially thoughts, cannot be reduced to sensory-motor processes and that we must look elsewhere for an explanation. One parallel but important stream of research demonstrates the importance of cognitive and social contributors to the sense of self-recognition.^66–70 These studies have developed independently of research on sensory-motor processes but share the idea that intentions are important for retrospective inferences about self-recognition. They provide an alternative perspective on mental phenomena that are not easily accommodated by sensory-motor constructs, although it should be noted that some studies have not supported the proposal that intentions are necessary for representing the self.^45,64 A reexamination of the role of intention might therefore provide important clues to the abnormalities underlying FRS.

Finally, while the processes underlying self-recognition are slowly unraveling, the brain mechanisms contributing to the subjective experience that some behaviors are under external control still remain to be identified. These are core experiences in schizophrenia and must remain an active area of inquiry.

Funding

National Health and Medical Research Council Australian Research Training Fellowship (404117 to F.A.V.W.).