The spectrum of indications for ultralong-term EEG monitoring

Purpose: We assessed clinical cases to investigate the spectrum of indications for ultra-longterm EEG monitoring using a subcutaneous implantable device in adult patients with focal epilepsy. Methods: Electronic charts were reviewed from patients undergoing ultra-longterm recordings at the European Epilepsy centers Barcelona, Freiburg and Vienna. Specific patient settings approached in the three centers were analyzed, and the main clinical question was extracted. Results from recordings were analyzed based on the specific results and information obtained. Results: 24 patients in whom ultra-longterm recordings were available were analyzed. A total of 11 main indications for subcutaneous long-term EEG recordings were identified, including the identification of active epilepsy in patients with low seizure frequency, under-and overreporting of patients, differentiation of non-epileptic from epileptic events, assessment of seizure severity, circadian and multidian rhythms of seizure occurrence, validation of treatment efficacy, improvement of patient-based reporting and medicolegal evidence for seizure freedom. This is reported with patient-specific case vignettes. Conclusion: Ultra-longterm monitoring using subcutaneous implantable EEG devices can provide relevant diagnostic and treatment information in a large spectrum of clinical situations. This is discussed considering the intrinsic limitations of the method related to spatial coverage, sensitivity and validity as a biomarker of ongoing seizures.


Introduction
Imprecise documentation of epileptic seizures based on patientbased reporting has been described in several inpatient-and outpatient investigations [1][2][3][4][5][6].The lack of precision of patient reports affects semiological description and seizure classification based thereon [7] as well as underreporting and overreporting of seizure frequency [8,4].As diagnostic decisions as to the classification of seizures as epileptic or non-epileptic, therapeutic decisions as to medical or non-medical treatment strategies and medicolegal evaluations as to driving and working capabilities have so far been based on patients' reporting, the question arises as to whether new technologies which have recently become available for objective seizures monitoring [9,10] may improve this situation.
Here, we propose a scheme of potential indications for the use of one such objective approach, the application of ultra-longterm monitoring using subcutaneous EEG (subQ) recordings.Indications were identified in patient cohorts treated at three European Epilepsy Reference centers (Barcelona, Vienna, and Freiburg), and casuistic illustration of potential clinical usefulness is given based on experiences in patients undergoing subcutaneous ultra-longterm recordings at these sites.

Methods
Ultra-longterm EEG recordings are recordings extending over a period of several months.Indications for ultra-longterm EEG recordings using a unilateral subQ device implantation were identified by retrospective analysis of cases studied at the three epilepsy centers Barcelona, Vienna and Freiburg.The subQ-device used at the three centers (24/7 EEG™ SubQ, UNEEG medical A/S, Allerød, Denmark), consists of 3-contact electrode implanted subcutaneously covering the temporal lobe or other lobes as indicated in the figures below, EEG data are continuously transmitted to an external device and transmitted to a cloud; data were assessed using an automated algorithm first [11] and verified visually by certified clinical neurophysiologists in each case.A total of 24 cases of subQ-ultralong-term recordings in adult epilepsy patients performed in the period 2021-2024 were reviewed regarding peculiar questions related to the individual patient's clinical situation.
To ascertain the capture of ictal EEG patterns, all patients reported had continuous video-EEG monitoring with seizure recordings, and the orientation of the subcutaneous electrode was chosen to cover the seizure onset zone.Correlation of subcutaneous and scalp recordings were checked during simultaneous recordings.
Indications were separated, and patient histories were reduced to concise features relevant to the respective indications for each clinical indication identified.

Diagnostic scheme of clinical indications for ultra-longterm monitoring and individual examples of outcomes
In a review of clinical applications of subcutaneous ultra-longterm recordings of 3-12 months in patients at the three centers, the following indications were identified (Table 1): Ultra-longterm monitoring may serve to address several clinical questions in one given patient.E.g. diagnostic confirmation of an active epilepsy will have therapeutic consequences, may lead to the conclusion of high seizure-associated risks and a need for the use of intensified, timely targeted therapy, or additional monitoring, and may have medicolegal consequences.Yet, frequently one aspect is key for the clinical decision to perform monitoring, as given below in case reports.

Diagnosis of active epilepsy in patients with low seizure frequencies (A-2)
Epilepsy, according to the ILAE, is diagnosed based on events considered epileptic seizures based on patient histories and on additional indicators for an assessment of seizure recurrence.This is usually based on standard or 24-h-EEG recordings and MR imaging.In unclear cases, inpatient video-EEG monitoring is used to identify ictal EEG patterns and assure their correlation with electrographic signatures.
However, the duration of inpatient monitoring is limited to 1-2 weeks at most epilepsy centers, as is the duration of home videomonitoring [12-14,].This limits the diagnostic yield of recording seizures and ictal electrographic patterns and may thus cause problems to ascertain epilepsy in patients with low seizure frequencies who escape the capture of seizures in up to 30 % of patients [13,15].Ultra-longterm monitoring can identify ictal electrographic events also in patients with low seizure frequency and thus establish the diagnosis of epilepsy.
Case 1: Evidence for seizures occurring at low frequency This 34-year-old female started to have epileptic seizures at the age of 11, having suffered from herpes encephalitis at the age of 10.She was amnesic for most of her seizures, except for some peri-ictal meaningless actions.Her relatives described episodes with loss of consciousness and oral automatisms lasting 3-5 min, with an estimated seizure frequency of up to five per month and rare progression into a bilateral tonic-clonic seizure.MR imaging showed postencephalitic alterations with bilateral insular and temporal gliosis as well as right hippocampal sclerosis.Video-EEG monitoring during presurgical evaluation captured seizures with a right temporal and frontal rhythmical theta activity in all cases, none of which were remembered by the patient.Interictal spikes occurred either right or left temporal.Due to the extended bilateral brain lesions and risk of herpes reactivation after neurosurgical interventions, neither invasive video-EEG monitoring nor surgery was performed, and she continued medical treatment.Seizure reporting was unclear, and subQ EEG was implanted right temporal to obtain information on ongoing epileptic seizures.During one year of seizure monitoring, a total of 16 electrographic seizures were captured by subQ EEG (Fig. 1) with circadian predominance between 12 am and 3 pm, at a frequency which evades inpatient assessment.

Diagnostic classification of episodic signs and symptoms as epileptic or non-epileptic (A-3)
Patient-based seizure reporting is frequently ambiguous as to an epileptic or non-epileptic nature.This is particularly difficult to judge in patients with ascertained epilepsy.Objective monitoring of EEGcorrelates of seizures can contribute to solving this issue.CASE 2 A 31-year-old right-handed individual presenting with non-lesional focal epilepsy originating in the left posterior quadrant, possibly adjacent to the language area, underwent evaluation.Initial assessment, facilitated by a translator due to language constraints, disclosed a history of experiencing two to three seizures weekly.Subsequent videoelectroencephalography (VEEG) monitoring revealed besides a

Table 1
Clinical indications for use of ultra-longterm EEG recordings.posterior focal epilepsy left-sided, concomitant psychogenic episodes lacking electrical correlates, posing a challenge for the patient to distinguish between the two phenomena.Semiology of epileptic seizures started without an aura and was marked by a loss of awareness, a fixed gaze, and later onset of oral and manual automatisms which correlated rhythmic theta activity in temporo-parietal region left.In contrast, psychogenic non-epileptic events (PNEE) were characterized by bradylalia and generalized weakness of the extremities, without EEG correlate.

Major Indication category
After subcutaneous EEG (SubQ-EEG) posterior left-sided implantation, the patient self-reported 12 seizures in their diary, while automated SubQ-EEG analysis identified 31 events.However, visual scrutiny concluded all instances as masticatory artifacts.Consequently, the patient's classification as drug refractory is revoked, as the episodes reported are confirmed to be non-epileptic, and the focal epilepsy is effectively managed with antiseizure medications (ASMs) (Fig. 2).

Assumed incorrect reporting of epileptic seizure frequency (A-4)
In patients with retrograde amnesia to seizures, nocturnal seizures, and seizures occurring unwitnessed.
Uncertainty of patients about seizure occurrence can lead to both underreporting and overreporting of seizures.Underreporting can be assumed in patients with retrograde amnesia, and overreporting can occur if patients interpret unspecific signs as possible indicators of seizures.Overreporting has recently been reported to occur as frequently as underreporting [8], and either of them may lead to false diagnostic and therapeutic decisions, as exemplified below.
CASE 3 A 53-year-old right-handed female patient underwent a comprehensive evaluation at our institution.She was diagnosed with left posterior quadrant epilepsy (PQE) associated with a complex cortical malformation, presenting considerable challenges for surgical intervention.Her living circumstances, residing alone, coupled with a lack of seizure awareness, hindered the provision of reliable seizure frequency data.However, she suspected that she was having seizures on an almost daily basis.Brain MRI findings revealed left occipital schizencephaly along with polymicrogyric cortex, collectively indicative of a complex cortical developmental malformation.Video-electroencephalography (VEEG) monitoring identified seizures displaying temporal semiology with onset in the left posterior temporal region.Subsequent utilization of posterior left-sided subcutaneous EEG (subQ-EEG) aimed to ascertain the actual seizure burden to optimize treatment.
Follow-up assessments utilizing SubQEEG identified nine events occurring over 399 days (Figs.3a and 3b).This observation contrasted considerably with the patient's subjective perception, revealing a significantly lower seizure burden.Consequently, the suspected burden was reduced, and adequate seizure control was achieved through antiseizure medications (ASMs).As a result, invasive diagnostic procedures in potentially eloquent areas were temporarily postponed.

Assessment of potential overreporting of symptoms considered seizure-related
Patients frequently report symptoms or signs potentially indicative of a preceding seizure.Unspecific symptoms may, however, lead to an over-or completely false reporting of seizures.This case exemplifies this: Case 4 (A-4) This 42 year-old female started to experience episodes consisting of a feeling of déjà-vu and an ascending feeling of heat followed by paresthesia in her left face lasting up to 30 seconds at age 35 years.The episodes initially occurred once every one to two months but subsequently appeared in clusters with up to 25 episodes within 1-4 days recurring every one to two months.Routine EEG showed bitemporal independent interictal spikes.MRI was normal.The episodes did not respond to several antiseizure medications.During the first video-EEG monitoring performed at age 41, no habitual episodes could be captured.During a second video-EEG monitoring performed six months later, 18 habitual seizures consisting of the typical aura described above accompanied by oral automatisms with preserved awareness could be recorded.The patient noted all seizures.Ictal EEG showed a right temporal EEG seizure pattern.90 % of interictal spikes occurred over the right temporal lobe.Right temporal SubQ was implanted to assess the actual seizure frequency and, more specifically, the number of seizures versus other events during and apart from seizure clusters.At the first visit, three months after the implantation of SubQ the patient had entered five events in the diary and had noted an additional five events in her personal seizure diary (a total of 10 events).None of these events were accompanied by EEG seizure patterns on SubQ and thus were classified as non-epileptic Fig. 4a).
At the subsequent follow-up visit six weeks later, the patient had noted a cluster consisting of 20 events occurring within four days.19 of these events could be correlated with EEG seizure patterns.In conclusion, in this patient, SubQ was useful in objectively documenting seizure frequency, especially during seizure clusters, and preventing overreporting of non-epileptic events outside seizure clusters.It should be noted that focal seizures with retained awareness are not always reflected by changes on scalp-EEG, especially if EEG seizure activity originates and remains confined to mesial temporal structures.Similarly, exclusively mesial temporal ictal activity therefore might also be missed on SubQ.However, in our patient during scalp video-EEG monitoring all habitual seizures with preserved awareness were accompanied by ictal scalp EEG patterns.EEG seizures patterns on scalp and on SubQ were quite similar.During video-EEG monitoring we could not observe any ictal EEG patterns arising from other brain regions.Because SubQ was implanted over the right lateral temporal lobe, we can virtually exclude that any ictal activity was missed on SubQ.We are therefore confident that Sub Q was reliable in differentiating focal seizures with retained awareness from non-epileptic events (Fig. 4b).

Verification of treatment efficacy (B-5)
Like any information on seizure frequency and severity, the efficacy of new treatments (changes in antiseizure medication, reprogramming of a neurostimulation device or epilepsy surgery) is judged based on seizure diaries.In cases of underreporting or overreporting of seizures, this can lead to false conclusions concerning treatment efficacy.
Case 5 This is exemplified by a 52-year old female patient who had focal non-lesional epilepsy of right temporal origin as shown by inpatient video-EEG monitoring, showing concordant right temporal interictal spiking and right temporal seizure onset with scalp recordings.Epilepsy was resistant to four antiseizure medications, including polytherapy.Yet, she was unwilling to perform epilepsy surgery as she assumed to have a low seizure burden based on her memory of seizures.SubQ recordings were performed, as there was discordance of seizure reporting both during inpatient monitoring and in the outpatient situation with evidence from witnesses of seizures which were not reported in the seizure diary.SubQ recordings showed a much higher seizure occurrence rate of 6-9/month, leading to the add-on introduction of additional antiseizure medication.Whereas seizure frequency appeared low and unchanged in the seizure diary, subQ EEG provided evidence for a considerable seizure reduction under the add-on treatment (Fig. 5).Such divergences may lead to false treatment decisions and judgment as to seizure-associated risks.

Seizure risk assessment supporting treatment decisions (B-6)
Presently, the decision to reduce antiseizure medication is based on patients' reports on seizure frequency and clinical parameters, disregarding EEG recordings, at least in focal epilepsy.This is based on the limited value of focal interictal epileptic discharges as predictors of seizure occurrence.Ultralong-term monitoring introduces the option of including the assessment of ictal subclinical electrographic discharges in the decision process (Figs.6a and 6b).This may warn about the reduction of antiseizure medication, as exemplified below.
Case 6 A 38-year-old right-handed individual presenting with right posterior quadrant epilepsy (PQE) due to focal cortical dysplasia (FCD IIa).The patient underwent surgical intervention at our institution in 2020 involving the placement of invasive subdural electrodes.Subsequently, a resection was performed, including the right posterior temporal and inferior parietal regions.Several months post-surgery, the patient began manifesting episodes of novel semiology.Subsequent video electroencephalography (VEEG) monitoring revealed these episodes to be psychogenic non-epileptic events (PNEE).Concurrently, there were uncovered subclinical seizures (SCS) of which the patient lacked awareness.In response, the implantation of a central right SubQ-EEG device was undertaken to distinguish between these two forms of events and to guide therapeutic interventions accordingly.Subsequent follow-up revealed two significant developments: a reduction in the frequency of PNES following the initiation of psychotherapy and a concomitant increase in SCS.Consequently, adjustments to antiseizure medication (ASM) were warranted to manage epileptic activity.However, given the patient's lack of awareness regarding the SCS, there was initial resistance to modifying the ASM regimen, necessitating persuasion based on insights obtained from subQ-EEG data.

Assessment of circadian and multidian seizure cycles (C-7)
Patient-based seizure reporting frequently does not provide valid information on circadian and multidian fluctuations of seizure occurrence [6].Subcutaneous recordings provide objective data on both circadian and multidian seizure cycles of occurrence.This is depicted in two patient examples with a circadian preference for seizure occurrence and multidian rhythms of seizure occurrence (Figs.7a and 7b).
A circadian predominance of seizures in the early afternoon was identified based on subcutaneous recordings in patient1 reported above.
Case 7: Multidian seizure cycles This was a 45-year old male patient with epilepsy since the age of 7. Seizures consisted of a short rising epigastric sensation followed by confusion, oral and manual automatisms and speech difficulties for several minutes and occasional progression to bilateral tonic-clonic seizures.3T MRI showed right hippocampal sclerosis, and ictal EEG patterns consisted of right temporal rhythmical theta activity.Despite pharmacoresistance, the patient refused surgical therapy.Due to amnesia for almost all of his seizures, a SubQ EEG system was implanted over his right temporal region for a period of one year.SubQ EEG recordings newly provided evidence of a constant and precise periodicity in seizure occurrence with cyclical clustering every 3-4 weeks (Fig. 7b) and almost seizure-free periods in between, which was not inferable from the seizure diary.
Both circadian and multidien cycles of seizure occurrence may be valuable for steering the timing of treatment (chronotherapy) [16].

Assessment of seizure severity and resulting seizure-associated risks (C-8)
Patients frequently can only report the initial focal aware phase of a seizure and not its further progression or may have complete amnesia for their seizures.Nocturnal seizures and the presence of bilateral tonicclonic seizures are considered factors increasing the risk of SUDEP in patients with epilepsy [17].Subcutaneous EEG recordings allow to assess of seizure duration based on the presence of electrographic ictal patterns and/or associated characteristic muscle artifact patterns during a tonic or clonic phase of a seizure (Fig. 8).This information can contribute to decisions on the need for treatment escalation.

Improving patient-based seizure documentation based on subQrecording-based feedback (D-9)
As stated above, patients can be uncertain as to the interpretation of subjective feelings as potential indicators of epileptic seizures.Informing patients about the results of objective documentation of their seizures may contribute to educating them to better categorize these perceptions as seizure indicators and improve seizure documentation in their diary.
Case 9 Improved seizure documentation with feedback from subQ recordings occurred in the case of a 50-year old female patient with right temporal seizures origin due to multiple cavernomas since the age of 42 years.She had focal unaware seizures with mild semiology consisting of staring, unresponsiveness, oral and manual automatisms, behavioral disturbance and at least partial retrograde amnesia for seizures.
MRI revealed multiple cavernomas infra-and supratentorial, two of them near the right hippocampus.During an in-hospital video-EEG monitoring, more than 60 short seizures with impaired awareness were captured, whereas the patient did not notice any of them.All documented clinical seizures were associated with right temporal rhythmical ictal patterns.Due to multiple cavernomas, a surgical intervention was considered to carry a high risk of bleeding.She thus preferred to continue pharmacological treatment; efficacy was, however, difficult to assess due to her inability to document seizure frequency.During six months of subQ EEG recordings, she received feedback on the number of seizures objectively documented.This led to an approximation of reported and documented seizures over time, reducing the difference between patient reporting and objective assessment of ictal electrographic seizures during the subsequent three-month period, even if time stamps were not always identical (Fig. 9).
Improving patient-based seizure awareness has been reported using micro-phenomenological training of patients to recognize early ictal signs better [18].EEG recordings, as performed here, may be another approach, possibly providing confirmatory information on interpreting ambiguous symptoms perceived by patients during or after seizures as valid indicators of seizures.Increase patient awareness of seizures may secondarily have positive aspects on compliance with medication intake.

Medicolegal decisions related to assurance of seizure freedom (E-11)
Complete seizure freedom for a defined period of time (e.g., one year) is a critical criterion for several medicolegal decisions, particularly for authorization to carry out certain professional activities and the allowance to drive vehicles.
Case 10 This 37 years old male started to have seizures at age 28 years.The patient suffered from focal impaired awareness seizures, which were noticed only by the patient's wife, while the patient himself was amnesic for these events and even denied their existence.Three of these seizures were witnessed by the patient's wife during the first three years after the disease's onset.In addition, the patient experienced rare focal to bilateral tonic-clonic seizures with a frequency of one per year.Routine EEG showed right temporal interictal spikes.MRI was normal.During video-EEG monitoring performed at age 31 years, two focal seizures with impaired awareness were recorded consisting of staring and manual automatisms.Ictal EEG showed a right temporal seizure pattern, and interictal EEG showed frequent right temporal spikes.After a modification of antiseizure medication, the patient remained seizure-free, both from focal impaired awareness seizures and bilateral tonic-clonic seizures.After a seizure-free observation period of one year, the patient resumed driving.The patient works as a technical field service consultant and has to drive long distances for many hours a day regularly.However, after a seizure-free period of more than five years, an episode suggestive of a focal impaired awareness seizure took place while driving.The patient drove to a customer but, after arriving, was amnesic for the whole drive.No accident occurred.Because the patient's work Fig. 8. Transition from the tonic to the clonic phase of a seizure in subQ-EEG recordings in a patient with right temporal lobe epilepsy due to a cavernoma who was amnesic for the majority of his seizures.depends on driving, the patient agreed to a SubQ implantation to objectively document seizure freedom to allow further driving.Implantation was performed three months ago over the right temporal region.No seizures were recorded so far.In conclusion, SubQ provides an objective tool to assess seizure freedom in patients with rare focal impaired awareness seizures, especially if patients themselves are amnesic for seizures and need to drive for professional reasons.

Conclusions
The range of valuable indications for the use of ultralong-term subcutaneous EEG recordings is large.It extends from the diagnosis of epilepsy to seizure documentation, patient instruction, risk assessment, objective and possibly timely steering of treatment, objective evaluation of treatment efficacy (see also [19]), and medicolegal decisions.Ultra-longterm EEG recordings open a new window for using ictal electrographic discharges for patient management.This allows to address a broad spectrum of clinical questions that go far beyond the underlying fundamental question of whether seizures can be proven, as discussed by Pathmanathan et al. [20].
One aspect of additional information provided by ultra-longterm monitoring beyond the pure presence of seizures is objective information on their timing.This information may be used to adjust the timing of medical or neurostimulation treatments on a circadian or multidian basis, depending on individual patients' cycles of seizure occurrence (personalized chronotherapy).Identified periods of seizure clustering may be used for acute or subacute interventions (e.g.Fountain [16]).Furthermore, multidian cycles have been shown to considerably contribute to the performance of seizure forecasting algorithms, as shown in several recent publications [21][22][23][24][25][26].Seizure cycles, as reported by patients, are already used by a first algorithm used to indicate individual seizure probabilities, and objective data on seizure cycles may, in the near future, lead to improved forecasting, warning and personalized patient care [27,28].Furthermore, information on the timing and severity of seizures can objectify risks relevant to treatment decisions and in a medicolegal context.
Despite the limited spatial coverage of the subcutaneous electrodes used, ictal EEG patterns can be identified at temporal and extratemporal recording sites.Whereas standard EEG and interictal epileptic activity is useful in about 60 % of patients to support the diagnosis of epilepsy [29,30], the extension of recording duration to many months opens up a completely new window of EEG utilization based on the identification of ictal electrographic patterns in focal epilepsy.
The usefulness of subcutaneous recordings with the specific device requires the active participation of the patients for the continuous streaming of EEG data.Thus, patients' good adherence to data recordings is a prerequisite for valid interpreting the ultra-longterm data.

Limitations
Objective seizure assessment using subcutaneous EEG recordings is limited to seizures with clear electrographic ictal EEG patterns and requires knowledge of the topography of such patterns in the individual patient.It is known that in about 15 % of patients with focal epilepsy, even with the complete 10-20-electrode set, no clear ictal EEG patterns can be identified, and patients without ictal scalp EEG patterns may assumedly not be assessable using subcutaneous recordings either.The available 3-channel EEG subQ EEG system is certainly limited in its spatial coverage of the head, and patients with difficult-to-recognize scalp EEG patterns, with varied extension across brain regions, or with considerable muscle artifacts obscuring EEG patterns (e.g. during purely tonic motor seizures) may thus not be suitable candidates.

Final remarks
A remarkable discordance of patient-based diaries from results of subcutaneous recordings was not only observed in this study on indications of ultra-longterm EEG recordings but also in a number of previous publications [10,31].This raises several points for future discussion: 1) How should individual patients be selected in whom additional information from ultra-longterm EEG recordings contributes to improved clinical care?The degree of patient ability to correctly document seizures dramatically differs according to video-EEGbased assessment, and a personalized selection of devices will be needed, including some parameters given in the indication list, like nocturnal or unwitnessed seizures.2) Do all electrographic ictal events represent clinical seizures, and do all ictal EEG patterns not reported as seizures prove underreporting?So far, some characteristic features differentiating subclinical from clinical features have been assessed for temporal lobe seizures [32]; correct separation of subclinical vs. clinical EEG patterns was possible in 75-80 %, leaving open the interpretation of up to 25 % EEG patterns.Additional research on extratemporal patterns, simultaneous subcutaneous recordings with video-EEG and databases to imply AI-based algorithms may improve this situation.3) Given the broad spectrum of clinical settings discussed, should ultralongterm monitoring be offered for improved patient care or to improve medicolegal decisions for all patients who may have a potential added value for their management?Additional evidence will be needed in larger patient cohorts to answer these questions, given the considerable workload for EEG analyses and the financial burden on the medical system.

Declaration of competing interest
Rodrigo Rocamora has received honoraria as a medical advisor and for lectures from UNEEG.
Christoph Baumgartner has received personal honoraria for lectures or advice from UNEEG related to the topic.
Andreas Schulze-Bonhage has received research funding from UNEEG and from the Epilepsy Foundation and personal honoraria for lectures or advice from UNEEG related to the topic.
The other authors have no potential conflicts of interest to disclose.

1 )Fig. 1 .
Fig. 1.SubQ-EEG based documentation of rare seizure occurrence (1.3/month), as indicated by blue dots/circles representing electrographic seizures.Note that 12 am here is indicating noontime and 12 pm midnight.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 2 .
Fig. 2. Patient report showing marks of PNES during subQ-EEG monitoring.The placement of the subcutaneous electrode (left temporo-parietal) is indicated on the left-sided head scheme.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3b .
Fig. 3b.Case 3 report showed improvement of electrographic seizures over time on subQ-EEG.The placement of the subcutaneous electrode (left occipital) is indicated on the left head scheme.Electrode contacts: D: distal, C: center, P: proximal.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 4b .Fig. 5 .
Fig. 4b.Results for SubQ for the first 4.5 months after implantation.During the second observation period, the patient noted a cluster of 20 events occurring within four days.19 could be correlated with EEG seizure patterns.Note that 12 am here is indicating noontime and 12 pm midnight.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 6a .
Fig. 6a.Typical subclinical seizure pattern in patient 6. Electrode contacts: D: distal, C: center, P: proximal.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 9 .
Fig. 9. Feedback on objective seizure frequency was provided since month 2 of recordings.Note that deviations from reporting to EEG-based monthly seizure counts became less expressed during the following three months of recordings with information on objective seizure numbers.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)