Sense of control, selective attention and cognitive inhibition in pediatric functional seizures: A prospective case-control study

Abstract

Purpose:

To date, laboratory-based experimental behavioral methods have not been used to identify factors associated with pediatric functional seizures (FS), leaving a critical gap for effective treatment development.

Methods:

Children ages 13–18 with video-EEG-confirmed FS were matched to controls (MCs) based on income, sex, race, and age. A modified Stroop task which included a condition requiring participants to report the ink colors in which seizure symptom words were written (e.g., “shaking” in blue) measured selective attention and cognitive inhibition through response time. The magic and turbulence task assessed sense of control in three conditions (magic, lag, turbulence). Children with FS were asked to report premonitory symptoms predicting FS.

Results:

Participants included 26 children with FS and 26 MCs (Mean_age=15.2, 74% female, 59% white). On Stroop, children with FS had a slower reaction time (Mean=1193.83) than MCs (Mean=949.26, p=0.022) for seizure symptom words. Children with FS had significantly poorer sense of control in the turbulence condition of the magic and turbulence task (Mean=−3.99, SD=8.83) than MCs (Mean=−11.51, SD=7.87; t(20)= −2.61, p=0.017). Children with FS (Mean=−1.80, SD=6.54) also had significantly poorer sense of control in the magic condition than MCs (Mean=−5.57, SD=6.01; p=0.028). Ninety-eight percent of patients endorsed premonitory symptoms.

Conclusion:

Compared with MCs, children with FS have 1) poorer selective attention and cognitive inhibition when presented with seizure-related information and 2) lower sense of control (i.e. poorer awareness that their control was manipulated). Premonitory symptoms were common. Sense of control, selective attention, and inhibition may be novel treatment targets for FS intervention.

INTRODUCTION

Functional seizures (FS) are a subtype of functional neurological disorder (FND; sometimes referred to as psychogenic non-epileptic seizures or PNES) characterized by seizure-like symptoms without associated epileptiform activity.¹ Incidence of FND in pediatric neurological services is at least 6 per 100,000 in children under 16,² and approximately 9.8 per 100,000 in children ages 15–19.³ Further, about 20% of patients in seizure clinics are eventually diagnosed with FS.⁴ FS are debilitating to patients and families, affecting physical functioning, academics, and peer relationships.^5–7

There is significant heterogeneity in the development of FS, and biopsychosocial models are used to conceptualize the etiology.^8–10 However, although onset is typically in adolescence or early adulthood,¹¹ most research on FS has focused on adults with these data often generalized and applied to children with FS. This creates a significant gap in the literature, as the predisposing factors for FS appear to differ between adults and children.¹² For example, while physical and sexual abuse are common in adults with FS, they have not been found to be significant risk factors for pediatric FS.^13–15 Thus, there is a critical need to consider novel factors specifically related to pediatric FS.

Studies have identified several novel risk factors for FND in adults including decreased sense of control and abnormal attentional focus. Sense of control is a single dimension in overall agency and is defined as the extent to which a person perceives they are in control.¹⁶ Research in adults with FND has assessed sense of control using two paradigms: the Libet’s clock task and action-effect binding.^17–20 In the Libet’s clock task, a red ball moves around an unnumbered clock face and participants push a button to stop the ball after a random amount of time.²¹ Participants then move the ball back to 1) where they first intended to move their finger to push the button or 2) to when they actually pushed the button. In this task, patients with FND, including FS, were found to have delayed self-report of intention to move as compared with controls.^{17, 18} Results from a second task, action-effect binding, demonstrated that when voluntary actions were followed by a tone, indicating the action led to an effect, healthy controls reported that their actions and the tone occurred closer together as compared with adults with FNDs.²⁰ Further, findings in adult and pediatric populations have demonstrated both functional²² and structural^23–25 abnormalities in areas of the brain related to movement and sense of control.²⁶ (e.g., corticospinal tracts or temporo-parietal junction).²² These findings indicate differences in sense of control between adults with FND and healthy controls. In pediatric populations, children with FS report no control over FS symptoms,²⁷ indicating they perceive FS to be involuntary.

Regarding attention, FND symptoms increase when attention is placed on the symptoms and improve with distraction.²⁸ One study demonstrated impairment in habituation among adults with FND compared to those with anxiety, which was explained as the result of poor selective attention.²⁹ Another study demonstrated abnormal attentional focus in adults with functional tremor during movement and improvements in functional tremors with distraction.³⁰ Children with FND have been found to have decreased selective attention via increased errors on a modified Stroop task as well as impairments in attention switching and inhibition as measured by the attention switching and go/no go tasks respectively.³¹

Overall, the roles of sense of control, selective attention, and inhibition in the generation and maintenance of FS in children remain unclear. Further, the majority of FS research uses clinical measures to assess neuropsychological factors specific to FS in this age group and there is no research using laboratory-based experimental behavioral testing methods in this area. The aim of this study is to evaluate sense of control, selective attention, and cognitive inhibition in relation to pediatric FS. We hypothesized that patients with FS would have poorer selective attention and inhibition, and have poorer awareness that their control is manipulated as compared to MCs.

METHODS

Design Overview

In a 1:1 matched study design in children with FS and matched healthy controls (MCs), each participant completed a laboratory visit involving assessments of sense of control, selective attention, and cognitive inhibition. Additionally, children with FS reported the occurrence of premonitory symptoms, or symptoms preceding FS.

Participants

Prospective participant enrollment was from November 2016 to March 2020 at the University of Alabama at Birmingham (UAB) with participants with FS enrolled consecutively and MCs identified after a participant with FS was enrolled and tested. Institutional Review Board (IRB) approval and written informed consent and assent were obtained prior to enrollment. Participants with FS were invited to join the study upon referral to Dr. Fobian’s Functional Neurological Disorder Clinic at UAB. Eligibility criteria for children with FS included ages 13–18 and video-EEG-confirmed diagnosis of FS. Comorbid epilepsy was only accepted if patient’s neurologist confirmed 6-month absence of epileptic seizures (6-month seizure freedom mirrored local adult studies where this time is required for return to driving). Exclusion criteria for individuals with FS and MCs included substance use, psychosis, or severe intellectual disability as assessed via parent report.

Controls were recruited via social media and fliers hung in appropriate community locations (i.e., schools, local YMCAs) and matched to those with FS on age, sex, race, and household income. Sex and race required an exact match. Age was matched ±1 year, and income was matched ±1 of the following intervals: 1) Below $20,000, 2) $20,000–39,999, 3) $40,000–59,999, 4) $60,000–79,999, 5) $80,000–99,999, and 6) Above $100,000. Additional exclusion criteria for MCs included any parent-reported mental health or medical diagnoses.

Measures

Demographics and Premonitory Symptoms.

Age, sex, race, income, and medications were reported by parents. Children with FS also retrospectively reported premonitory symptoms preceding their FS via a question that asked, “If you have symptoms before your FS which tell you that you are about to have a FS, please indicate them below.”

Sense of Control.

Sense of control was measured by the magic and turbulence task. This computer task has been used to assess sense of control in a variety of conditions and across the lifespan.^32–35 The task consists of two phases: 1) a game phase in which participants use the mouse to move a cursor along a horizontal track at the bottom of the screen to catch the falling X’s and miss the falling O’s (Figure 1A), and 2) a judgment phase in which participants judge how much control they believe they had (control judgment) and how well they believe they performed (performance judgment) during that condition. In addition to this normal condition, three conditions manipulate control in the game: 1) the lag condition changes the cursor movement to be delayed relative to the speed at which the participant moves the mouse, 2) the turbulence condition moves the cursor side to side unrelated to the movement of the mouse, and 3) the magic condition indicates X’s have been caught if the cursor is within 10 pixels of the X, even if the cursor does not touch the X. The conditions are presented in a randomized order, each occurring six times, and average scores of control and performance are computed for each condition based on participant judgments. Sense of control is then assessed by calculating three summary control scores. These scores represent a comparison of the participants’ sense of control over and above their perception of their overt performance in conditions in which their control is manipulated (magic, turbulence and lag conditions). To calculate this, we computed 3 summary control scores, namely, the contrast: (performance judgmentN – control judgmentN) – (performance judgmentM – control judgmentM) where the N refers to the normal condition and M refers to the manipulated condition (magic, turbulence or lag).³³ Greater negative numbers indicate greater sense of control (i.e. better understanding their control is being manipulated).

Figure 1.

Screenshots of the Magic and Turbulence and modified Stroop Tasks

Selective Attention and Cognitive Inhibition.

Participants completed a modified computerized Stroop task which assesses selective attention (i.e., the ability to focus on a particular stimulus of one’s choice) and cognitive inhibition (i.e., the ability to suppress attention to unwanted stimuli).^{36, 37} The original version of the Stroop task requires individuals to read a word written in a specific color (red, blue, or green) and to report the color in which the word is written. The modified version used in this study had 3 conditions: 1) color words congruent with the color in which they appear (e.g., “red” written in red ink), 2) color words incongruent with the color in which they appear (e.g., “red” written in blue ink) and 3) words describing symptoms of epileptic seizures written in a colored ink (e.g., “unconscious” written in green ink; Figure 1B). The third condition was added to assess selective attention and cognitive inhibition in direct relation to symptoms of epileptic seizures. Conditions were presented in the following order: congruent, incongruent, seizure symptoms. The participants completed each condition twice, with the first time used as practice to ensure understanding of the task and minimize the novelty effect of seeing seizure symptom words for the first time. For each condition, reaction time was measured as the average amount of time in milliseconds it took for the participant to identify the color in which each of the words was written. Greater reaction time indicated a participant had increased difficulty inhibiting attention to extraneous stimuli (i.e., incongruent color words or seizure symptoms) and thus had poorer selective attention to the focal stimuli. Error rates for each condition were also assessed. These rates were dichotomized for the final analysis due to non-normal distributions, which were not able to be corrected through transformations.

Verbal Intelligence Quotient.

Participants were given the Shipley-2 questionnaire as a measure of verbal IQ.³⁸ The Shipley-2 includes 40 terms for which participants must identify the word from 4 choices that is most like the original term (e.g. “Divest”, answer choices: “dispossess,” “intrude,” “rally,” or “pledge”). The Shipley-2 is scored by referencing age-based norms. This was used as a covariate in Stroop analyses due to the emphasis on reading during the task.

Power Analysis

A power analysis for a repeated measures ANOVA within factors analysis was conducted in G*Power (v3.1). Using a within groups effect size of f=0.7 based on previous literature for the Stroop and magic and turbulence tasks,^{34, 37} a desired 95% power, a correlation of repeated measures of 0.5, two groups, and two measurements, this analysis indicated a required minimum of 10 participants to assess differences in sense of control and selective attention between teens with FS and matched controls.

Data Analysis

Normal distribution of variables was assessed. Sphericity was assessed for all ANCOVA analyses. If the assumption of sphericity was violated, Greenhouse-Geisser correction was used based on Epsilon values <0.75. No other issues of normality were noted. Missing data were assumed to be missing completely at random, and complete case analysis was used to assess the relationships between variables.

First, two repeated measures ANOVAs investigated within-subjects differences in reaction time among the 3 Stroop conditions for both children with FS and MCs. Next, three separate Cox’s Regressions were used to assess differences between patients with FS and their matched healthy controls on errors in the Stroop task. Then three additional repeated measures ANCOVAs assessed differences in reaction time between children with FS and MCs for each Stroop condition (matched, unmatched, and seizure symptoms) while controlling for verbal intelligence. Finally, three paired samples t-tests investigated group differences in sense of control on the three magic and turbulence task conditions (turbulence, lag and magic).

RESULTS

There were 52 participants who completed the Stroop and the magic and turbulence tasks: 26 with FS and 26 MCs (Mean_age=15.4, 86% female, 67% white). However, some data from the Stroop and magic and turbulence tasks were missing due to computer malfunction or user error, resulting in samples sizes of 23 with FS and 23 MCs for the Stroop task and 21 with FS and 21 MCs for the magic and turbulence task. There was no difference in number of participants taking psychotropic medications between children with FS and MCs. As premonitory symptoms were assessed in participants prior to the addition of the Stroop and Magic and Turbulence computer tasks, the sample sizes vary (see screening and enrollment data in Supplementary Figure 1). Premonitory symptoms were assessed in 42 children with FS (Mean_age=15.2, 74% female, 59% white). There were no significant differences in sociodemographic factors between participants who only reported premonitory symptoms and those who completed the Stroop and magic and turbulence tasks.

Of the 42 children with FS who completed the premonitory symptom assessment, all but one (98%) reported at least one premonitory symptom signifying an impending FS (Table 1). The most common premonitory symptom was headache (57%). Participants reported an average of 3.9 (SD=2.6) premonitory symptoms.

Table 1.

Demographics for children with FS (controls matched on age, sex, race and household income) and premonitory symptoms experienced prior to pediatric FS

Characteristics		%
Sex	Female	85.7
Race	White	66.7
	Black	22.3
	Other	16.6
Income	Below $20,000	16.2
	$20,000–$39,999	26.7
	$40,000–$59,999	3.2
	$60,000–$79,999	23.3
	$80,000–$99,999	18.2
	Above $100,000	12.4
		Mean ± SD
Age (yrs)		15.4 ± 3.0
Premonitory Symptom		# Reporting (%)
Headache		24 (57%)
Increased Heart Rate		19 (45%)
Blurry/Double Vision		16 (38%)
Mood Change		14 (33%)
Stiffness		13 (31%)
Pain		13 (31%)
Fatigue		12 (29%)
Nausea		10 (24%)
Aura		8 (19%)
Auditory Symptoms		8 (19%)
Dizziness/Disorientation		8 (19%)
Temperature Change (hot or cold)		6 (14%)
Skin Sensitivity		3 (7%)
Numbness		2 (5%)
Other		8 (19%)
Total reporting at least 1 premonitory symptom		41 (98%)
Mean (SD) # of premonitory symptoms reported		3.9 (2.6)

For children with FS, there were significant differences in reaction time among Stroop conditions, F(2,40)=7.91, p=0.001 (Table 2A). Post hoc comparisons indicated longer reaction time for both the incongruent condition (Mean=1307.42, SE=104.41; p=0.004) and the seizure symptom words condition (Mean=1193.83, SE=83.32; p=0.042) compared to the congruent condition (Mean=968.18, SE=97.06). There were no other differences between groups (Figure 2). For MCs, there were differences in reaction time among Stroop conditions, F(1.31,26.20)=37.32, p<0.001 (Table 2A). Post hoc comparisons indicated significantly longer reaction time in the incongruent condition (Mean=1342.19, SE=77.62) compared with both the congruent (Mean=856.98, SE=48.67; p<0.001) and the seizure symptom words conditions (Mean=949.26, SE=48.47; p<0.001). There were no other differences between groups (Figure 2).

Table 2.

A. Average Stroop condition reaction time (in milliseconds) and test statistics for ANOVAs comparing reaction time of adolescents across Stroop conditions for patients with FS and matched controls. B. Average Stroop condition reaction time (in milliseconds) and test statistics for ANOCVAs comparing reaction time of adolescents with FS and matched controls, adjusting for difference in Shipley Verbal IQ Score. Higher numbers signify slower reaction time.

A.
Group	Congruent Words	Incongruent Words	Seizure Words	F	p
FS (n=21)	968.2	1307.4	1193.8	7.91	.00**
Matched Controls (n=21)	857.0	1342.2	949.3	37.32	.00**

B.
Condition	FS (n=21)	Matched Controls (n=21)	F	p
Congruent Words	968.2	857.0	0.84	.37
Incongruent Words	1307.4	1342.2	0.18	.67
Seizure Words	1193.8	949.3	6.22	.02*

^*p≤0.05;

^**p≤0.01

Figure 2.

Mean reaction time by condition for children with functional seizures and matched controls and p-values indicating significance among reaction times.

*p≤0.05; **p≤0.01, ***p<0.001

For children with FS and MCs, there were no significant differences in errors for the congruent, incongruent, or seizure symptom words conditions (ps > 0.05). However, children with FS had longer reaction time for the seizure symptoms condition than MCs, F(1,19)=6.22, p=0.022. No significant differences were noted when comparing reaction times of children with FS and MCs for the congruent (F(1,19)=.84, p=0.37) or incongruent conditions (F(1,19)=.18, p=0.674; Table 3B, Figure 2).

Table 3.

Average magic and turbulence condition scores and test statistics for paired samples t-tests comparing FS and matched controls, N=46 (23 with FS, 23 MCs). Numbers closer to zero indicate poorer sense of control.

Condition	FS	Matched Controls	t	p
Turbulence	−4.0	−11.5	−2.61	.02*
Magic	−1.8	−5.6	−2.35	.03*
Lag	−5.5	−8.8	−1.21	.24

^*p≤0.05;

^**p≤0.01

A paired samples t-test revealed children with FS (M=−3.99, SD=8.83) had poorer sense of control in the turbulence condition of the magic and turbulence task compared to MCs (Mean=−11.51, SD=7.87; t(20)=−2.61, p=0.017; Table 3). Children with FS (Mean=−1.80, SD=6.54) also had poorer sense of control in the magic condition compared to MCs (Mean=−5.57, SD=6.01; t(21)=−2.35, p=0.028). No differences were found in the lag condition (t(19)=−1.21, p=0.24).

DISCUSSION

This is the first study that used laboratory-based experimental measures to assess differences in sense of control, selective attention, and cognitive inhibition between children with FS and matched controls. Results of this case-control study reveal three significant points. First, children with FS make less accurate judgments of sense of control compared to MCs. Second, children with FS have poorer selective attention and inhibition associated with information related to seizure symptoms. Third, almost all children with FS experience premonitory symptoms prior to FS.

In assessing sense of control, children with FS were less aware that their control was being manipulated during the magic and turbulence task. This occurred when control was both impeded (turbulence condition) and enhanced (magic condition). Although differences in sense of control have been demonstrated in the adult FND literature via fMRI²² and two computer tasks, an action effect binding task and the Libet’s clock task,^17–19 this is the first study to confirm this relationship in pediatric FS. This is a novel finding that distinguishes FS from MCs that may act as a changeable target in the treatment or prevention of FS.

Regarding decreased selective attention and inhibition, children with FS had a longer reaction time when presented with seizure symptom words compared to MCs. This indicates that children with FS were unable to selectively attend to the ink color and inhibit attention to the seizure symptom words, despite their irrelevance to the task, resulting in increased reaction time. The incongruent condition (i.e. color words presented in a text color different than the written word, e.g., “red” written in blue) would be expected to produce the greatest difficulty with selective attention due to the contextual similarity between the written words and the text color. However, children with FS performed similarly on the seizure symptom words condition and the incongruent condition, despite reaction time similar to MCs on the congruent and incongruent conditions. Thus, while overall selective attention and inhibition are not different in pediatric FS, children with FS had poorer performance in these areas during exposure to seizure related information. This finding differs from previous research highlighting elevated errors on the Stroop task and impaired performance on other measures of attention in pediatric patients with FND.³¹ However, the patient population in this previous study included children with various FND symptom presentations, including FS, and the study was conducted while patients had active motor-sensory symptoms.³¹ In some cases, these active symptoms (e.g. tremor) may have impacted testing performance above and beyond what was observed in patients with FS who had no acute symptoms during testing. Further, differences in the specific assessment outcomes (i.e. Stroop errors versus Stroop reaction time) and dimensions of attention (e.g. attention switching and sustained attention versus selective attention) being evaluated may also have contributed to these differing results. Alternatively, selective attention in children with FS may be different than for those with other FND symptoms.

Overall, the observed difficulty inhibiting attention to seizure symptom words in children with FS may contribute to the development of a stronger seizure scaffold.¹¹ Similar to a classically conditioned reflex, a seizure scaffold includes a series of perceptions and motor responses developed from previous experiences, which can be triggered by a range of internal or external stimuli.^{9, 11} Alternatively, children with FS may have more difficulty inhibiting interference from seizure symptom words because they are currently experiencing similar symptoms. Comparing children with FS to those with epilepsy might help determine if this is the result of having similar symptoms or a correlate of FS. However, our results are consistent with research demonstrating decreased habituation to tones in adults with FND, which was attributed to differences in selective attention²⁹ and build upon current literature regarding attentional differences in children with FND.

Finally, the majority of children in this study reported at least one premonitory symptom preceding their FS. Premonitory symptoms have been previously reported in children with FS³⁹ and noted to be beneficial as a target within treatment.^{40, 41} While most studies have focused on the objective semiology of pediatric FS, this is the first study, to our knowledge, to report the proportion of children experiencing premonitory symptoms and assess the frequency of specific, individual premonitory symptoms experienced by children prior to FS. The high incidence of premonitory symptoms linked to hyperventilation (e.g., tingling, numbness, muscle stiffness, dizziness, changes in vision) and autonomic arousal (e.g., increased heart rate, nausea, temperature changes) reported in this study^{42, 43} is consistent with past research. Specifically, studies have shown high rates of premonitory symptoms in children with FS when their carbon dioxide levels were below 30mmHg during a paced breathing task³⁹ and the occurrence of hyperventilation and hyperactivation of the autonomic nervous system in some children prior to FS.^{39, 44} This could suggest that elevated rates of hyperventilation and autonomic arousal may underpin some of the premonitory symptoms reported in this study. While only one participant reported changes in breathing (included as “Other” in Table 1), endorsement of the above symptoms may still indicate unrecognized hyperventilation, as past research highlights discrepancies in self-report and objective assessment of hyperventilation with patients often endorsing the above symptoms during hyperventilation.⁴³ However, not all reported premonitory symptoms are characterized by autonomic arousal or hyperventilation, such as the 19% of patients who reported auditory changes prior to FS. Thus, more research is needed to clarify the causal factors related to premonitory symptoms and if this differs among children with FS. Further, because children and their parents may not be aware of hyperventilation in relation to FS, this research should evaluate hyperventilation and autonomic arousal via both objective assessment (e.g., respiratory rate per minute or percutaneous CO₂) and self-report.

It has been suggested that adults with FS may sometimes choose to have a FS in order to relieve the discomfort of premonitory symptoms.⁴⁵ This is also a theory of Tourette’s disorder in which premonitory symptoms are explained as an aversive stimulus toward which tics are purposively directed and which result from poorer inhibition of attention to physical symptoms.⁴⁶ This theory may be consistent with the differences in selective attention and inhibition demonstrated by children with FS when exposed to seizure symptom words. However, this theory does not explain why individuals would respond with seizure-like symptoms to relieve premonitory symptoms. Therefore, it is more likely that relief of premonitory symptoms helps to maintain FS as opposed to being the etiological mechanism by which they are initially produced, and this may be exacerbated by poor selective attention, difficulty inhibiting attention when exposed to physical symptoms, and/or hyperarousal of the autonomic nervous system.

The above findings have significant implications for the treatment of pediatric FS. As many children with FS do not have psychiatric disorders prior to or after FS onset,⁴⁷ targeting mood and stress is unlikely to be an effective treatment for FS. Alternatively, sense of control, selective attention, and inhibition are promising novel treatment targets for FS. Further, the presence of premonitory symptoms in most patients provides additional potential for novel treatment approaches for pediatric FS, highlighting a unique opportunity for intervention prior to the progression of episodes. Given that relief of uncomfortable premonitory symptoms may reinforce the occurrence of FS, working to combat these symptoms may be effective in treating FS. Past research has highlighted acute improvements in symptom presentation through regulation of the autonomic nervous system and regulation of breathing.³⁸ Additional improvements may be made through 1) improving selective attention to inhibit attention to the premonitory symptoms and 2) increasing sense of control during the premonitory period. A recent randomized controlled trial (RCT) assessing Retraining and Control Therapy (ReACT), a cognitive-behaviorally based treatment for pediatric FS, suggests this approach may be effective.⁴⁸ In ReACT, patients develop a plan which aims to increase sense of control by using opposing responses during and after the premonitory period to prevent or stop FS symptoms and reduce catastrophic symptom expectations. ReACT significantly reduced FS compared to supportive therapy, and 82% remained FS-free for at least 60-days after treatment and 57% remaining FS-free 1-year after treatment,⁴⁹ suggesting targeting sense of control may be effective in treating FS.⁴⁸ However, the target mechanisms by which ReACT is effective are not confirmed.

Finally, the prospect of developing and disseminating a short-term, targeted treatment for FS would have many benefits. First, a brief, targeted treatment may increase the confidence of mental health care providers in treating FS.⁵⁰ Second, despite research demonstrating no significant improvement in FS by treating mood,⁵¹ providers often still recommend targeting mood to improve FS based on the traditional psychodynamic explanation for FND.⁵² However, treatment of mood often requires frequent, prolonged psychotherapy. Given the limited number of therapists comfortable treating FS,⁵⁰ a brief, targeted treatment for FS that does not target mood could expand the availability of FS-informed therapy and improve access to treatment. This is consistent with the biopsychosocial understanding of FS and with the advances in pediatric FS treatments which have demonstrated symptom improvements in some patients following the provision of psychoeducation on FS, use of habit reversal and biofeedback, consultation with schools, physical therapy, and other clinical targets outside of mood.^{41, 48, 53–57} Continued advancements would allow for further specificity regarding early treatment targets. As prognosis is related to duration of symptoms,⁵⁸ targeted treatment may lead to faster recovery and mitigate the negative effects of FS on school, work and social activities for patients and families.

This study also highlights the need for several new directions of research. First, differences in sense of control, selective attention and inhibition have not been previously demonstrated in pediatric FS, and their role as potential novel targets to more effectively treat FS in children must be assessed.⁵⁹ Second, additional controlled studies are needed using laboratory-based experimental behavioral evaluations (in addition to clinical assessments), fMRI or other biological or physiological measures to assess potential novel predisposing factors. These studies should include control groups such as siblings, children with similar symptom presentations (e.g. epilepsy) or children matched on demographics and psychiatric comorbidities. Third, assessing selective attention and inhibition in pediatric FS compared to epilepsy will be helpful to confirm if our findings are unique to FS or due to the similarity in symptoms they are experiencing. Finally, additional research can further assess premonitory symptoms in FS, building on current knowledge regarding 1) frequency, 2) etiology and relation to autonomic arousal, and 3) utility in treatments to prevent FS.

This study is the first controlled study to match children with FS on age, sex, race and household income and use valid experimental behavioral methods to assess neuropsychological differences in sense of control, selective attention, and inhibition. The sample size is adequately powered to detect significant differences between the groups, and it assesses novel factors related to pediatric FND. Limitations include a cross-sectional design, use of verbal IQ instead of a more comprehensive IQ measure, and no control for psychiatric diagnoses or symptom similarity (e.g. epilepsy). Further, while exclusion criteria included comorbid medical and mental health diagnoses and severe intellectual disability, attention deficit/hyperactivity disorder was not directly assessed, which could affect attention outcomes on the Stroop task. However, as participants with FS and MCs performed similarly on the congruent and incongruent conditions of the Stroop task, we do not believe this explains the differences in the seizure word condition. Finally, the majority of participants identified as white and female; however, this is expected given the demographic make-up typically observed in FND patient samples.⁶⁰

Conclusions

This study is the first to demonstrate impairments in sense of control, selective attention, and cognitive inhibition in pediatric FS. Overall, this study highlights potential novel target mechanisms for FS interventions in children. Additionally, the presence of premonitory symptoms predicting FS provides an opportunity for intervention to prevent the onset of FS episodes.

Highlights.

• Children with functional seizures demonstrated poorer selective attention and cognitive inhibition as compared with matched controls when presented with seizure-related stimuli.

• Children with functional seizures had poorer awareness that their control was being manipulated.

• A majority of children with functional seizures reported experiencing premonitory symptoms prior to functional seizures.