Introduction

Psychogenic nonepileptic seizures (PNES) are paroxysms of altered sensory, cognitive and/or motor manifestations, with or without alteration in consciousness that may resemble epileptic seizures (ES), but do not originate from abnormal electrical brain activity. Most presentations of PNES fulfill criteria for Functional Neurological Symptom Disorder (Conversion Disorder – CD) under Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5),¹ reflecting a seizure type in which psychological conflicts are converted to somatic ES-like symptoms. The classic psychological process is described as, when focusing on external seizure manifestations (or other somatic symptoms), the patient’s own internal stressors (e.g., the weights of self-culpability surrounding psychological conflicts and/or self-responsibility to resolve them) are alleviated from conscious awareness. ²

Several etiological frameworks exist for CD of seizure type. No single mechanism or contributing factor has been known to be necessary and sufficient to explain all heterogeneous cases of PNES or any other CD manifestation. ³ One framework conceives PNES as arising from a biopsychosocial, multifactorial etiologic model as illustrated in Figure 10.1. PNES are conceptualized to evolve from serial interacting factors over time, comprised of Predisposing, Precipitating and Perpetuating factors.^{4, 5} Predisposing factors consist of inherent or imbued elements within the patient’s constitution, which confer vulnerability toward eventual development of PNES later in life. Precipitating events occur in individual’s lives which patients may identify as “the cause” (e.g., a head injury, an assault, a blast exposure). Perpetuating factors, both internal and external, aggravate the underlying conflicts/issues. Whereas one factor may appear to play predominant roles in a particular patient, this model stipulates that other important factors may also be involved and should not be overlooked.⁶

Figure 10.1 Biopsychosocial etiological model: Development of psychogenic nonepileptic seizures via multifactorial contributions including Predisposing, Precipitating and Perpetuating factors.

From Reuber, with permission.⁶

Studies have reported a notably high co-occurrence rate of PNES and traumatic brain injury (TBI), with a weighted average of 43% among 1,039 adult patients with PNES across 17 studies.⁷ While PNES have an estimated prevalence of 2–33/100,000 (more in women)⁸ and a history of TBI is endorsed by 12% of the general population in developed countries (mostly men),⁹ it is unlikely that coincidental co-occurrence can entirely explain such high comorbidity of PNES and TBI. In applying the biopsychosocial model, this review discusses how TBI may subserve contributing roles as Predisposing, Precipitating and Perpetuating factors in the development of PNES.

Structural or Functional Effects of TBI as Predisposing/Precipitating Factors

Earlier investigations have shown that in cases where PNES were preceded by TBI, 87% of such cases occur with mild TBI (mTBI),¹⁰ while 13–33% were with moderate to severe TBI.^{10, 11} Both of these cited studies determined TBI severity using criteria as suggested by Annegers et al.,¹² characterizing TBI as mild when the associated loss of consciousness or posttraumatic amnesia was less than 30 minutes in duration. Patients with longer duration of posttraumatic unconsciousness or amnesia, the presence of skull or radiographic evidence of TBI (based on conventional imaging modalities) were characterized as having moderate to severe TBI. These studies could be superficially interpreted as mild TBI (mTBI), being somehow more pathogenically associated with PNES. However, since PNES have been reported following TBIs of all severity levels (mild, moderate and severe),^{10, 11} the observed distribution (87%) may simply mirror the overall incidence rate of mTBI, representing 80–90% of all head injury cases in the general population.¹³ Within the past decade, significant advances in neuroimaging methods have uncovered macro- and submacroscopic abnormalities in patients with all severity levels of TBI (including mTBI) and/or PNES – findings which were previously not evident with conventional imaging modalities. Pertinent findings postulating the contribution of TBI to the neurobiological underpinnings of PNES will be further explored in this review.

Nascent literature pertaining to neuroimaging of PNES has shown divergent results, with different studies identifying varying potential regions of interest despite having study subjects from a similar population.¹⁴ In view of these divergent results, this review will highlight notable regional substrates, as well as more widespread network abnormalities where key studies have converged to postulate the neurobiological overlap between TBI and PNES.

The uncinate fasciculus (UF) is a white matter association tract that connects anterior temporal structures (e.g., amygdala, hippocampus) to prefrontal cortices (e.g., orbitofrontal cortex), and it has been directly implicated in psychiatric disorders.¹⁵ More specifically, the temporo-amygdala-orbitofrontal network has been hypothesized to be critically involved in integration of emotional states with cognition and behavior.¹⁶ Notably, a diffusion tensor imaging (DTI) study showed that the UF is affected in 29% of 34 patients who sustained mTBI.¹⁷ Applying DTI methods to patients with patients with PNES, one study observed abnormal DTI parameter (fractional anisotropy – FA) involving the left UF, superior temporal gyrus and subcortical structures when compared to healthy controls.¹⁸ Another DTI study found similar FA values between patients with PNES versus controls, but highlighted that patients with PNES demonstrated a significant asymmetry in the number of reconstructed UF streamlines (right greater than left). This asymmetry in UF streamlines was not exhibited in controls.¹⁹ De novo PNES has been reported in case series after general neurosurgery²⁰ and systematically investigated after epilepsy surgery, affecting 2.4–8.8% of patients who had neurosurgery.^21–23 Most of epilepsy surgery cases entailed complete or partial resections for temporal lobe, frequently interrupting the UF and its limbic connections.

Among the major functional limbic networks, the UF exhibits high heritability as measured by degree of genetic variances between monozygotic and dizygotic twins.²⁴ The UF has also been characterized as having a protracted maturation process, undergoing developmental evolution well into the third decade of life.²⁵ In the aforementioned DTI study, an inverse correlation was observed between the degree of UF asymmetry and age at PNES onset (i.e., higher asymmetry indices were associated with younger ages at PNES onset).¹⁸ A corollary to these observations could be that TBI at an earlier age (combined with inherited UF disturbance, when present) may contribute in parts to microstructural damage to the UF that confer vulnerability to later development of PNES. As such, TBI and genetic constitution could be considered as predisposing factors in the biopsychosocial etiologic model for PNES.

The default mode network (DMN) represents a brain network composed of the following nodes: rostral anterior cingulate gyrus, superior temporal and supramarginal gyrus, posterior cingulate gyrus and ventromedial prefrontal cortex.²⁶ The DMN subserves self-reflective mental activity and is more active at rest than during tasks requiring attention. Disruption of functional interactions within DMN after TBI was associated with impairment of both self-awareness²⁷ and inhibitory cognitive control.²⁸ A study utilizing high-density electroencephalography (EEG) found that PNES frequency correlated with degree of hypo-synchronization within the prefrontal and parietal regions, both of which are integral components of the DMN.²⁹

A novel Integrative Cognitive Model (ICM) aims to bring together existing theories of PNES phenomenology toward a singular explanatory framework.³⁰ The ICM postulates that dysfunction of inhibitory processing represents a key late factor among a series of steps that precipitate PNES. This impairment of inhibitory control may potentially arise from disturbance within the DMN.³¹ As such, TBI insulting critical brain substrates (e.g., DMN) could contribute in part as Precipitating factors in the biopsychosocial etiologic framework for PNES.

Recent studies of patients with PNES have utilized robust data-driven, functional connectivity analyses which investigate whole brain networks rather the pre-defined regions of interest (e.g., independent component analysis, graph theory). These advanced methods have demonstrated that in addition to changes within the DMN, more widespread alterations in fronto-parietal (attentional), emotional processing, executive control and sensorimotor networks.³² Rather than a focal neuroanatomical process, these and other functional connectivity studies³³ advocate conceptualizing PNES as having disrupted neural networks, with seizure manifestations developing from alterations within, as well as abnormal interaction across, the aforementioned networks (Figure 10.2). Distinct disruptions to various parts of these involved networks may also contribute to the observed variations in PNES phenomenology.³⁴ These neural networks are thought to represent recursive large-scale networks, typical of those involved with conscious awareness of sensory input and behavior (metacognitive processes).³¹ Some have posited that metacognition is a unique process to humans, which may explain the lack of animal models of conversion disorder.³⁵ The long-range trajectory of the connectivity within such networks augments susceptibility to axonal shearing or straining forces from TBI, reflecting another plausible neuropathophysiological link between TBI and the development of dissociative symptoms accompanying PNES.⁷

Figure 10.2 Conceptualization of psychogenic nonepileptic seizures as neural network disorder: Seizure manifestations develop from alterations within as well as abnormal interaction across key networks that contribute to altered self-reflection, dissociation, impaired action-selection and altered perception/somesthesis.

From van der Kruijs et al.,³¹ with permission.

Overall, the present literature in the neuroimaging of PNES is limited but growing, with early studies generally of small sample sizes that did not compare patients with PNES to controls matched for neuropsychiatric conditions. More studies comparing the overlap of network abnormalities evident in PNES with neuroanatomically based brain disorders (i.e., TBI) are needed. Further understanding of the neural basis behind the contribution of TBI to PNES does not negate or contradict the more commonly cited cognitive-behavioral mechanisms, but rather underscores the inextricable concurrence of multiple factors (structural, functional and psychological) as stipulated in the biopsychosocial framework for PNES. The psychological effects of TBI will be explored next.

Psychological Effects of TBI as Precipitating Factors

TBI, even at a mild degree, can induce significant level of stress and anxiety. About 14% of subjects who sustained mTBI from motor vehicle accident were diagnosed with acute stress disorder within 1 month of the injury.³⁶ Patients who sustained mTBI were likely to have ongoing symptoms at 12 months, compared to patients with orthopedic limb injuries.³⁷

In predisposed individuals (including 80% of patients who experience acute stress disorder), stressors from the trauma can induce long-term changes in stress-responsivity and metacognitive function, ultimately fulfilling the diagnostic criteria for posttraumatic stress disorder (PTSD).³⁸ Moreover, several studies involving either civilians or veterans have highlighted the significant association of PTSD with the development of PNES.^{2, 39, 40} Among US Veterans, PTSD was shown to significantly increase the likelihood of diagnosing PNES versus epileptic seizures when mTBI was the purported seizure etiology.⁴⁰ Based on these observations, investigators have inferred that PTSD may subserve a moderating role in the development of PNES, without which mTBI might be one-step away or entirely unrelated to PNES⁷ (Figure 10.3). PTSD may also mediate the development of other somatic symptoms in Veterans who have sustained mTBI, including cognitive⁴¹ and severity of postconcussive⁴² symptoms.

Figure 10.3 Mild TBI leading to PNES via several potential pathways. In predisposed individuals, trauma can induce long-term changes in stress responsitivity and metacognition that define PTSD. In turn, PTSD could subserve a moderating role toward the development of PNES. Illness beliefs combined with anxious misattribution of acute symptoms (i.e., post-concussive symptoms) can foment maladaptive symptom modelling. Abbreviations: PNES, psychogenic nonepileptic seizures; TBI, traumatic brain injury; PTSD, postraumatic stress disorder.

Modified from Popkirov et al.,⁷ with permission.

Among patients who develop PTSD, some symptoms usually emerge within the first 3 months after the trauma. A subsequent delay of months or years (rarer, delayed expression PTSD) can transpire before DSM-5 criteria for the diagnosis are met.¹ Moreover, some investigators have conceptualized PNES as a manifestation of the dissociative subtype of PTSD.⁴³ Altogether, the expression of PTSD following TBI could be interpreted as a Precipitating factor in the biopsychosocial etiologic model for PNES.

Mild TBI frequently combines highly emotionally stressful situations (e.g., vehicular accidents and their aftermath, combat theater-related events) with acute neuropsychiatric dysfunction in motor/sensory faculties, memory and consciousness (e.g., post-concussive syndrome). Under such scenarios of heightened distress, illness beliefs regarding risk for posttraumatic epilepsy, combined with anxious misattribution of post-concussive symptoms can foment maladaptive symptom modelling – patient’s own learnt mental representation of seizure symptoms and expectation of how seizures express⁴⁴ (Figure 10.3). When confronted with trauma reminders, unbearable stressors or unconscious emotional/somatic cues, PNES are triggered as part of a maladjusted emotional and physical stress response.^{7, 45} Similar cognitive behavioral mechanisms have been proposed to instigate other non-PNES functional neurological disorders. In a systemic review of 869 patients with motor or sensory conversion symptoms, 37% had reported preceding physical injury.⁴⁶ Among 50 patients with psychogenic (functional) movement disorders (PMD), 22% experienced preceding physical injury typically involving the peripheral soft tissues.⁴⁷ In many of these reported cases, the demonstration of functional symptoms following brain-sparing traumatic injuries robustly support the cognitively and emotionally mediated effects of the physical traumas, independent of any neuronal disturbances within the brain.

An aforementioned study reported that 80% of patients reported the physical event within 3 months prior to onset of the PMD.⁴⁸ In addition, since maladaptive symptom modelling following TBI may subserve a role in the expression and triggering of PNES, it could be considered as a Precipitating factor in the biopsychosocial etiologic model for PNES.