Functional Stroke Symptoms: A Narrative Review and Conceptual Model

Illness experiences are framed by cultural and social contexts (24). Historically, functional symptoms were associated with psychological trauma and stress, and for a proportion of patients, these are important risk factors. A systematic review and meta-analysis of case-control studies reported an association between functional motor disorder (FMD) and maltreatment (25). Likewise, prospective studies have reported that functional neurological disorders (FND) have stronger associations with childhood sexual abuse and physical neglect than do organic disorders (26–28). Experiences of sexual abuse are more common (though not invariable) in patients with FND, compared with patients diagnosed as having depression and healthy control subjects, and a greater proportion of patients diagnosed as having FND report severe life events in the month prior to symptom onset (29). Certain events may be more common in these patients—for example, experiences of emotional neglect, bullying, and other interpersonal problems (25–27, 30). Changes to the FND diagnostic criteria reflect findings that stress, abuse, or trauma are sufficient but not necessary for the occurrence of functional symptoms (31). Supporting this change are findings from a recent study using medical records from a large psychiatric organization in London, which reported similar rates of childhood sexual abuse in FND patients and those with general psychiatric diagnoses (27). Therefore, although a thorough patient history is always important, evidence of past trauma or psychiatric comorbidity is not necessary for the diagnosis of functional stroke symptoms.

A review of evidence suggests that experiences of organic illness can influence later functional symptoms (32). A high proportion of FMD patients and patients with functional seizures have a history of organic neurological disorders, physical injury, or surgery (6, 33), and injury to the affected limb has been found to be associated with acute onset of functional weakness (34). A systematic review of functional motor and sensory symptoms reported physical injury in 37% of patients prior to symptom onset (35). It is noteworthy that injury was more likely to be reported prior to functional paraparesis (35), which might indicate that different triggers are relevant to lateralized or hemiplegic presentations, as seen in FSMs. Analyses of data from a national birth cohort study showed that poor parental health during adolescence was longitudinally associated with medically unexplained symptoms occurring in adulthood (36). This research indicated a role for symptom modeling in FND etiology, which we expect applies to functional stroke.

Exposure to stroke or cardiovascular disease, in immediate social circles or in oneself, may predispose individuals to functional stroke symptoms, because a lay understanding of bodily function informs presenting symptomology and interpretation. Hypothetically, a history of stroke may bias a patient’s response to harmless sensory symptoms, which could contribute to the occurrence of functional stroke symptoms. Another likely predisposing factor is the presence of known stroke risk factors, such as hypertension (11, 19, 32). Individuals (and physicians) may be primed to expect stroke symptoms if told that their risk of stroke is increased and may misinterpret bodily sensations accordingly (see the section below on cognitive biases). A contemporary model outlines how illness experiences and expectations are neurally encoded and then key to the generation of functional symptoms, which may explain why particular symptoms occur in an individual patient (37). Exposure to stroke-related illness needs to be assessed in FSMs.

The wider social context may also act as a predisposing factor for functional stroke. Löwe and Gerloff (20) suggested that cultural beliefs are a vulnerability factor in the development of functional symptoms and use of health care services. In a consecutive case series in a Middle Eastern hospital, patients of Arab and African ethnicity reported twice the rate of functional stroke, compared with those who were Eastern, Asian, or Western, although this effect may also be partly explained by socioeconomic status (38).

Public health campaigns that aim to raise awareness of stroke symptoms have been shown to result in increased visits to stroke websites, calls to helplines, visits to emergency departments for stroke symptoms, and rates of thrombolysis treatment (39). Although such effects may be short term (40), it is plausible that such campaigns influence presentations of functional stroke symptoms. Careful prospective controlled studies of such campaigns are needed to test this hypothesis scientifically.

Biological stress pathways are implicated as risk factors for functional symptoms. Aybek et al. (41) reported a correlation between life events, cortisol, and α-amylase levels in FND patients but not in control subjects. Furthermore, FND patients have higher baseline cortisol and α-amylase levels, compared with healthy control subjects (41), suggesting chronic activation of biological stress pathways, which may precede symptom onset. Activation of stress pathways may therefore have a similar role in predisposing individuals to functional stroke symptoms, although evidence is currently limited.

Across the functional disorder literature, there is a growing consensus that in addition to the social and biological factors mentioned, patients are predisposed to functional symptoms by premorbid cognitive biases (37, 42, 43) and aberrant emotion processing (44). In their integrative conceptual model of functional seizures, Reuber and Brown (45) suggested that preexisting mental representations of symptoms form a “scaffold” for functional symptoms, which is activated when symptoms occur. We suggest that a similar process occurs in functional stroke, where, for example, a memory of a past ischemic event may evoke a pattern of stroke-like symptoms. However, in this review, we emphasize the role of cognitive and emotional processes as perpetuating factors that become salient after the occurrence of symptoms, rather than predisposing symptom onset (Figure 1).

Compared with patients with depression or healthy control subjects, patients with functional diagnoses report more life events prior to symptom onset (29). Keynejad et al. (31) suggested that patients with high biological propensity for functional symptoms may be at risk following a relatively minor stressor, compared with individuals with low biological susceptibility, who may be at risk only after exposure to a severe stressor. In addition to acting as predisposing risk factors, physical illnesses may act as a precipitating event for a proportion of FSMs. In interviews, 80% of FMD patients reported physical events, such as illness, injury, or infection, in the preceding 3 months, and physical injury to the symptomatic limb was reported by 20% of patients with sudden-onset weakness (12, 33). Such events are acute stressors in themselves, and they also influence the interpretation of symptoms, potentially increasing the likelihood of perceiving symptoms as threatening.

Initial stroke-like symptoms are likely a result of physiological arousal, which is misinterpreted and drives further physical symptoms (Figure 1). In retrospective interviews, 59% of patients with sudden-onset functional weakness reported symptoms of panic (34). In another retrospective study, 38% of FMD patients reported a panic attack at the onset of symptoms (33). Persons experiencing such symptoms would likely believe that they were indicative of a serious medical disease, particularly given increasing public knowledge about stroke. Anxiety triggers further autonomic system arousal, adding to initial symptoms. This leads to the testable hypothesis that anxiety symptoms are a more frequent accompaniment to the onset of functional stroke symptoms, compared with vascular stroke and potentially other functional presentations.

Compartmentalization, a form of dissociation, may also play a role in the emergence of functional symptoms. In this account, symptoms result from disturbances in processes underlying consciousness and self-control, such as memory retrieval. Applying this theory, stroke-like symptoms arise from “the retrieval of inappropriate perceptual hypotheses from memory during the creation of primary (mental) representations” (46). For example, unexplained weakness could arise as a result of the automatic memory selection of a movement task leading to inappropriate motor action.

A significant proportion of research emphasizes the role of emotion processing in functional disorder etiology. Compared with healthy control subjects, patients with functional seizures experience difficulties in emotion identification, selection and implementation of regulation strategies, and accurately appraising external emotion information (58). A systematic review of experimental studies provided neurobiological support for these findings; during emotion experimental tasks, functional symptoms were found to be associated with increased limbic and motor region activation and aberrant prefrontal and paralimbic activation functionally connected to motor areas (44). Compared with healthy control subjects, FND patients showed increased activity and connectivity between the amygdala and supplementary motor area during the recall of stressful events or when responding to negative emotions or emotive faces (59, 60). Subsequently, Pick et al. (44) suggested that aberrant emotion processing mediates associations between environmental risk factors and functional symptoms. Moreover, such aberrant neural activity has shown associations with symptom duration and severity (59), indicating a neurophysiological pathway for persistent functional symptoms. Overall, findings suggest that emotion-related processes play a key role in functional symptom maintenance and form an important treatment target, but this has not been explored in FSMs.

The three-systems model (TSM) provides a framework to understand the mechanism through which functional stroke symptoms are reactive to external events via emotional behavior (61–64). The TSM suggests that individual emotional responses reflect the unique pattern of three discordant components: subjective cognitive states, physical arousal, and behavioral responses, with each component having a different loading (65). We postulate that individuals with functional stroke symptoms demonstrate an emotional response pattern loaded heavily on physiological components. As such, functional stroke symptoms can be conceptualized as autonomic reactions to a life event or social context, with varying degrees of concordance from cognitive (e.g., worry, catastrophizing) and behavioral (e.g., avoidance, help seeking) components. The TSM can account for the heterogeneity of patients with functional symptoms. Those whose emotional responses are dominated by physiological and behavioral components may be more likely to interpret symptoms as a deterioration in physical health, because they do not experience corresponding subjective cognitions (e.g., “I feel anxious”). Patients with concordant cognitive components may be more likely to attribute symptoms to stress or emotional states. These differences have important implications for treatment (66).

As symptoms continue and interfere with daily life, classical conditioning offers an explanation for how stress and functional symptoms can become coupled (32, 61). Physiological responses to stress (e.g., somatic symptoms) can become a conditioned response, and over time, the threshold for the physiological response falls until symptoms occur in response to minor stressors or even the memory of an event (32).

Stroke-like symptoms are understandably associated with a set of acute responses from patients, caregivers, and medical health professionals concerned about the possibility of serious cerebrovascular events. In the immediate aftermath of functional stroke symptoms, patients may experience intense medical attention, emergency hospital admission, and extensive physical assessments and treatment (13). Our systematic review (9) and previous studies (38) have reported that rates of functional symptoms are higher in acute medical services, compared with community settings, and are identified to a greater extent in samples receiving thrombolysis, compared with samples not receiving thrombolysis. Learning theory can provide an explanation for how such responses perpetuate symptoms. Symptom-related beliefs and behaviors are easily and quickly internalized as they are associated with positive reinforcers and negative consequences (67). Urgent responses from health professionals reinforce the perception of symptoms as serious and threatening and induce further anxiety. Care and attention received from family and friends can act as a positive reinforcer (68, 69). There is some evidence that the belief that stress is the cause of the symptoms is more common among the relatives of patients with FND than the patients themselves, meaning that relatives may encourage the patient to get more rest (70). It is also possible that to avoid further upsetting the patient, caregivers support patients’ views and coping behaviors, even if they disagree. Relatives have been found to be more pessimistic regarding the potential duration of symptoms and the possible emotional impact of the symptoms (70). Levy et al. (67) reported that children whose parents had more anxious, attentive responses to somatic symptoms complained of more symptoms than did children whose parents gave less solicitous responses. Wives of men who had experienced mild stroke reported taking on a hypervigilant role, which can sustain anxiety around symptoms and prevent full recovery (71).

In response to ongoing symptoms and social reinforcement, patients can develop an illness identity that contributes to symptom maintenance (72). This could be relevant for patients with more chronic functional stroke symptoms. Strong illness identities have been found to be associated with greater disability and poorer psychological outcomes in other functional patient groups, making illness identity an important treatment consideration (73).

Following illness or injury, patients adapt their behavior to reduce the risk of further harm, but such a process can inhibit recovery. In a prospective study of anxiety after diagnosis of a stroke or transient ischemic attack, phobic anxiety disorders were the most common anxiety subtype (74). All anxious patients reported phobic avoidance in physical, social, and daily activities, and avoidance was associated with poorer quality of life, compared with patients who did not report avoidance (74). The associations between avoidance and functional symptoms suggest that avoidance is a perpetuating factor that reinforces illness-related beliefs and illness identity (75, 76). Observational studies have shown that patients admitted to a hospital for stroke spent almost 50% of time resting in bed and only 13% of time doing activities to support physical mobility (77). Inactivity before and after discharge can lead to problems such as postural hypotension, muscle atrophy, and loss of strength (78, 79). Caregivers may also contribute to inactivity when they assist patients with daily tasks or encourage rest. These issues are also relevant to functional stroke symptoms. Patients who experience unilateral weakness avoid using the affected side, even when symptoms have subsided, which leads to loss of strength. Subsequently, subjective weakness and soreness occur when the muscles are used, which reinforces beliefs of physical damage and increases attention to any weakness in the limbs. Similarly, patients may experience lightheadedness when active after prolonged rest, which can be perceived as a possible sign of stroke and can lead to further activity avoidance. Finally, patients may become more alert to unilateral symptoms, checking for facial asymmetry or being hypervigilant of any potential stroke-like symptoms.

Symptoms of physiological arousal, such as dizziness, palpitations, transient numbness, weakness, or fatigue, might ordinarily be normalized. However, in the context of particular predisposing and precipitating factors (i.e., illness-related experiences and significant life stressors), symptoms can be perceived as more sinister and can be attributed to physical injury or disease. Concurrently, increased attention is given to symptoms, and the reactions of caregivers and others reinforce beliefs that symptoms are indicative of a serious event. All these responses increase anxiety, which elicits further symptoms, help-seeking behavior, and other behavioral responses (45). In chronic cases, a cycle of symptoms, attention, anxiety, expectation, and cognitive and behavioral responses perpetuates symptoms and, in some patients, contributes to disability. It is likely that only parts of the suggested model (Figure 1) will apply to any individual patient, and during assessment, the model should be applied so that the patient is provided with an explanation that best reflects his or her experience.

Our model provides a preliminary theoretical framework with which to begin the design of a functional stroke care pathway, which could also benefit patients with stroke. Figure 2 shows a potential stepped-care pathway, similar to that previously recommended by the National Health Service Scotland (80), which we have adapted for functional stroke symptoms.

First, our model can be used as an accessible tool for clinicians in stroke services to enhance their knowledge of functional stroke and communicate a model of understanding symptoms to patients. Considerable advice is emerging that is designed to inform neurologists about how best to manage FND, and this needs to be adapted and disseminated among stroke staff who work in relatively unique neurology settings (81–83). When possible, they should show by demonstration (e.g., drift without pronation and Hoover’s sign) and by explanations that avoid the perception that functional symptoms are a “last resort” diagnosis (84). Stroke doctors may be the starting point of the multidisciplinary approach and must be able to give clear, unambiguous explanations that set the tone for later interventions (85), even when patients present with transient, subjective, or sensory symptoms that may be more difficult to demonstrate.

Following clear communication of symptoms, the perpetuating factors outlined in our model could be targeted by interventions with use of cognitive or behavioral techniques in primary care or hospital settings (86). In line with long-standing recommendations (85) and a recent consensus statement (87), these interventions would need to be adapted and specialized to functional symptoms and could be delivered by a range of health professionals, such as psychologists, behavior therapists, and physiotherapists, who have the necessary expertise. Drawing on Chalder and Willis’s (22) comprehensive discussion of transdiagnostic versus specific processes in functional disorders, we suggest that a functional stroke intervention should start by identifying the unique pattern of cognitive, emotional, and behavioral processes relevant to the etiology of a patient’s symptoms. For example, patients who demonstrate a bias for cognitive responses may be more responsive to cognitive techniques, whereas those with a propensity for physiologically oriented responses and behaviors may benefit more from desensitizing interventions aimed at reducing arousal (66). Taking a transdiagnostic approach allows our model to be applicable to both FSMs and those with functional symptoms in addition to a vascular stroke. However, this relies on clinicians discussing with patients that symptoms and disability in stroke are, as in all other conditions, not entirely attributable to physical pathology but are also affected by cognitive and behavioral responses. Cognitively and behaviorally informed approaches have supporting evidence in improving outcomes for patients with functional symptoms (87–89).

A small number of FSM patients may require higher-intensity and longer-term intervention with psychiatric or psychological services, a referral the stroke clinician should suggest in the discharge letter to the general practitioner. Case studies are useful references that provide detailed examples of how such treatments can be implemented with patients who have acute-onset functional symptoms (90). Our model posits that factors evident during assessment can determine the appropriate referral pathway for an individual patient; for example, anxiety, strong illness-related beliefs, and avoidance behavior are likely important treatment targets in any treatment pathway, whereas significant trauma history may require earlier input from more specialist services. Some evidence suggests that patients who have high levels of physical dysfunction may benefit from access to intensive, specialized physiotherapy and that this is especially effective if accessed during the acute onset of symptoms (91). Neurophysiotherapists in stroke or neurology wards are more likely to have the skills to provide this intervention, compared with clinicians on general medical wards or in community settings. Therefore, an impetus may exist to allow FSM patients to have a longer inpatient stay.