Ictal semiology in anterior cingulate epilepsy: A systematic review

Abstract

Seizures originating from the anterior cingulate cortex (ACC) have distinct clinical features but can be difficult to identify in frontal lobe epilepsy (FLE). This systematic review examines the key semiology of ACC seizures and their anatomical correlations. A systematic search was conducted following PRISMA guidelines, including studies reporting ictal semiology, invasive EEG findings, and surgical outcomes in patients with ACC seizures, allowing for the establishment of anatomical and clinical correlations with a high level of confidence. Studies based only on stimulations were excluded. We selected 23 studies including 93 patients (57% males, 46% children). MRI positive (59%), invasive monitoring (74%), surgery (100%) with Engel class I outcome (80%) were the main characteristics. Cortical lesions were found by histology in 86% of the surgical specimen, including focal cortical dysplasia in 67%. The level of confidence in epileptogenic zone (EZ) localization was considered high and very high in 87% of patients. Auras reported by 58% of them mostly included affective (fear or negative emotional feelings) and/or autonomic symptoms (80%). The main ictal signs consisted of facial emotional expressions (46%), autonomic features (48%), vocalization and sudden complex/hypermotor behavior (60%). Spectacular manifestations with preserved awareness, verbalizations with emotional content, laughter, ictal pouting (“chapeau de gendarme”) can also point to ACC involvement. In contrast, dystonic/tonic–clonic features, head and eye deviations were less frequently observed (<20%). Loss of consciousness was reported in 35% of patients. Immediate recovery at the seizure‐end was usual. Short seizures (<1 min) occurring in clusters during sleep were also characteristics. Combination of these features enhanced the likelihood of ACC origin. Interictal personality disorders which improved after seizure control can be also observed. ACC seizures are predominantly characterized by emotional, autonomic, and striking behavior manifestations contrasting with preserved awareness. These semiology markers support an anatomical and clinical entity and help to localize the EZ in FLE.

Key points.

• This systematic review based on 23 studies and 93 patients examines the key semiology of anterior cingulate cortex (ACC seizures) and their anatomical correlations.

• Auras were present in 58% of the patients, mostly affective (fear or negative emotional feelings) and/or autonomic.

• The main ictal signs consisted of facial emotional expressions (46%), autonomic features (48%), vocalization and sudden complex/hypermotor behavior (60%).

• Striking behavior manifestations contrasting with preserved awareness was highly suggestive of ACC involvement.

• Short seizures (<1 min) occurring in clusters during sleep were also characteristics.

• Dystonic/tonic‐clonic features, head and eye deviations were less frequently observed (<20%).

1. INTRODUCTION

Seizures originating in the anterior cingulate cortex (ACC) were first reported by Penfield and Jasper ¹ in 1954, who described them as “stereotyped acts linked to emotional or affective restraints,” reflecting “a continuum between erroneous visual perception, forced thoughts and acts, and complex motor behavior with prevailing emotional contents.” This description was later refined by Bancaud and Talairach, ² , ³ , ⁴ who defined a distinct clinical pattern, “the easiest to identify and to treat surgically, never seen in temporal lobe epilepsy.” They reported striking manifestations, including intense fright with facial expression of fear, shouts, screaming, vigorous and aggressive verbalizations, autonomic disturbances, and complex gesticulation. They noted that behavior could remain appropriate to the environment or appear strange, sometimes involving aggressive acts. Consciousness was disturbed but contact with the external world could be preserved and integrated into complex behavior.

Contrasting with these distinctive presentations, ACC epilepsy was later considered rare, under‐recognized, and frequently misdiagnosed due to its heterogeneous semiology and deep location. Seizures involving the cingulate cortex were described as manifesting with a broad range of non‐specific symptoms, making diagnosis and treatment challenging. ⁵ , ⁶ However, the same authors suggested that certain key clinical features may help identify ictal involvement of the ACC. Over the past decades, efforts have been made to better characterize frontal lobe epilepsy (FLE) and its related semiology. ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ Increasing evidence from stereo‐electroencephalography (SEEG) and successful epilepsy surgery has refined the classification of FLE subtypes along an antero‐posterior gradient, from prefrontal to precentral regions, and from mesial to lateral areas in adult and pediatric series. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ Studies focusing on the ACC have further detailed the semiology of seizures arising from this region. ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ Several case reports and small case series have highlighted illustrative features in both adults and children. ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ Moreover, distinct ictal signs such as fear, hypermotor behavior, and ictal pouting (chapeau de gendarme) have been analyzed for cortical localization, emphasizing ACC involvement. ³⁶ , ³⁷ , ³⁸ These findings have contributed to a more precise clinical characterization of ACC epilepsy. However, pure ACC seizures have been relatively rare in published studies, and terminology has been variably used in describing ictal features. In addition, the definition of ACC anatomical boundaries has evolved over time, and current subdivisions of the cingulate cortex ³⁹ , ⁴⁰ have not been consistently implemented.

To improve knowledge of ictal semiology in ACC seizures, we conducted a systematic review based on precise anatomical and clinical correlations. We investigated which symptoms and signs, whether isolated or clustered, were more frequently observed in ACC seizures than in other FLE subtypes and therefore may be suggestive of seizure origin in this region.

2. METHODS

2.1. Search strategy and eligibility criteria

This review was conducted according to PRISMA guidelines. ⁴¹ We systematically searched PubMed and Web of Science up to September 1, 2024, using the keywords: cingulate OR frontal lobe AND (epilepsy or seizure) AND (surgery OR outcome) AND (semiology OR video). Filters were set to English and abstract availability. We included original, peer‐reviewed research articles without time restrictions. Two independent reviewers (FC and ES) screened all titles and abstracts. Duplicates were manually removed. Full‐text eligibility was assessed independently by both reviewers using predefined criteria. Articles were eligible if they provided information on clinical features, invasive EEG recordings, epilepsy surgery with surgical outcomes, or brain imaging data. Studies were included only if they provided individual detailed seizure semiology with respect to ACC involvement. Case reports were considered if they reported precise anatomical and clinical correlations. Studies based only on brain stimulation findings were excluded.

2.2. Data extraction

For each selected study, we extracted the following variables: authors, year, number of reported patients, proportion of patients with the target EZ, gender, age, adult/children ratio, percentage of MRI positive cases, invasive EEG evaluations, operated patients, and favorable outcome according to Engel classification ⁴² (Engel class I with a minimum 1‐year follow‐up). Ictal semiology was analyzed based on the presence and type of aura (affective, autonomic, sensory, psychic, indescribable), vocalization or verbalization, laughter, autonomic signs, facial expression change, pouting (chapeau de gendarme), automatisms (motor gestural and oro‐alimentary), complex motor or hypermotor (hyperkinetic) behavior, eye and/or head deviation, dystonic posturing, elementary motor signs, bilateral tonic–clonic manifestations, loss of consciousness (LOC), and postictal confusion. The timing of each sign and symptom (early/late) during the seizure was noted if mentioned. Seizure duration and time of occurrence (day or night, during sleep) were recorded. Interictal behavior changes were also noted when reported. Each symptom or sign was classified according to the ILAE glossary. ⁴³ ACC localization was assessed based on the current subdivision of the cingulate cortex. ³⁹ , ⁴⁰

2.3. Risk of selection and assessment bias

Selection bias was evaluated by assessing whether studies enrolled a consecutive or random sample, clearly defined sampling methods, used a case–control design, and avoided inappropriate exclusions. Assessment bias was based on whether the interpretation of ictal semiology was conducted without knowledge of other data, particularly EEG and MRI findings. Risk was classified as low, high, or unclear. ⁴⁴

2.4. Reliability of the reference standard and level of evidence

Confidence in the reported EZ was graded into four levels (very high, high, moderate and low) based on MRI findings, invasive EEG data, and surgical seizure outcome (Engel class I, minimum 1‐year follow‐up). ⁴⁵ The overall level of evidence was classified as high, moderate, low, or very low. ⁴⁶

2.5. Statistical analysis

We performed a descriptive analysis of the patient population, with continuous variables reported as mean, standard deviation (SD), and range, while categorical variables were expressed as percentages.

One‐sided binomial test was applied to determine whether a symptom occurred in more than 33% of patients. Symptoms meeting this criterion were subsequently classified as typical. To assess the relative frequency of syndromes, we computed pairwise odds ratios and applied Holm correction to adjust for multiple comparisons. Statistical significance was established at p < 0.05.

3. RESULTS

3.1. PRISMA flow diagram

The PRISMA flow diagram (Figure 1) outlines the study selection process. We initially identified 894 citations, screened 183 abstracts, and reviewed 124 full‐text articles. A total of 23 studies were included, comprising a cohort of 93 patients.

FIGURE 1.

Flowchart of the systematic review.

3.2. Selected studies

We included 13 studies with at least two patients, ⁵ , ⁸ , ⁹ , ¹² , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ³⁴ , ³⁶ , ³⁷ , ³⁸ five of which specifically addressed cingulate cortex or ACC seizure semiology, ⁵ , ²⁰ , ²¹ , ²² , ²⁴ and 10 single case reports ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁵ (Table 1). Cohort studies had a mean sample size of 18 patients (SD: 15, range 2–57), with an average of four patients with the target EZ (SD: 6, range 2–23). Two studies from the same site and author were identified as having overlapping patient cohorts. ¹⁹ , ²¹ The study with the smaller sample size was excluded. ¹⁹ We analyzed clinical features based on ACC delineation, considering variability in posterior limits among the studies (Figure 2). In one study, ²⁴ we included patients with an EZ involving both the ACC and the anterior middle cingulate cortex (aMCC).

TABLE 1.

Characteristics of selected studies and patient population.

Study	Authors year	Focus of study	Nb patients	Nb of patients with target EZ	Mean age (years) min‐max Nb of children	Gender male %	MRI positive %/Nb	Intra‐cerebral EEG %/Nb	Operated %	Engel class I %/Nb	Level of confidence in EZ localization % %
A. Series including at least 2 patients
1	Pellicia et al. 24 2022	Cingulate epilepsy	57	23	5.8 (0–20) 20	57	48 11	78 18	100	100 23	Very high 100
2	Von Lehe et al. 20 2012	Cingulate surgery	22	19	35 (12–63) nd	68	84 16	68 13	100	63 12	Very high 63 High 16 Moderate/low 21
3	Souirti et al. 38 2014	CDG	36	10	31.2 (16–48) 0	80	50 5	90 9	100	100 10	Very high 100
4	Chou et al. 22 2020	Cingulate epilepsy	18	10	25.3 (11–40) 3	30	40 4	70 7	100	50 5	Very high 50 Moderate/low 50
5	Alkawadri et al. 21 2013	Cingulate epilepsy	14	5	28.4 (4–49) 2	80	100 5	0 0	100	60 3	Very high 60 High 40
6	Mihara et al. 8 1997	FLE semiology	18	3	31.3 (28–34) 0	0	66 2	100 3	100	66 2	Very high 66 High 33
7	Rheims et al. 37 2008	Hypermotor seizures	11	3	25.6 (10–34) 1	33	33 1	100 3	100	100 3	Very high 100
8	Clopenborg et al. 34 2021	Hyperkinetic seizures+fear	3	3	9.3 (6–11) 3	NR	0 0	100 3	100	100 3	Very high 100
9	Devinsky et al. 5 1995	ACC and behavior	2	2	42.5 (42–43) 0	100	100 2	100 2	100	50 1	Very high 50 Moderate 50
10	Lu et al. 23 2021	CDG	10	2	22 (20–25) 0	50	50 1	100 2	100	100 2	Very high 100
Total A			191	80	19.8 (0–63) 29	55	58 46	75 60	100	80 64	Very high 79 High 9 Moderate/low 12
B. Series including a single case with target topic
11	Bautista et al. 9 1998	FLE	9	1	34	1	1	0	100	100	Very high
12	Biraben et al. 36 2001	Fear	8	1	19	0	1	1	100	100	Very high
13	Sinclair et al. 12 2004	FLE	22	1	8	NR	1	1	100	100 a	High
Total B			39	3	20.3 (8–34) 1	‐	100 3	66 2	100	100 3	Very high High 33
C. Case reports
14	San Pedro et al. 25 2000	ACC/SPECT	1	1	40	1	0	100	100	0	Moderate
15	Seo et al. 26 2003	Pilomotor seizures	1	1	27	1	0	100	100	FU <1 year	Moderate
16	Mohamed et al. 27 2007	ACC/MEG	1	1	17	1	1	0	100	100	Very high
17	De Rose et al. 28 2009	Cingulate epilepsy	1	1	2.5	1	1	0	100	FU <1 year	High
18	Lacuey et al. 29 2015	ACC epilepsy	1	1	30	1	0	100	100	100	Very high
19	Mirandola et al. 30 2015	ACC epilepsy +PNES	1	1	13	0	1	0	100	100	Very high
20	Daniel and Perry 31 2016	Coprolalia	1	1	15	1	0	100	100	100	Very high
21	Koc et al. 32 2017	CDG	1	1	24	0	1	100	100	100	Very high
22	Qiao et al. 33 2017	Fear	1	1	22	0	1	100	100	100	Very high
23	Jayalakshmi et al. 35 2021	Smile and fear	1	1	4	1	1	100	100	100	Very high
Total C			10	10	19.5 (4–40) 4	70 7	60 6	70 7	100	70 7	Very high 70 High 10 Moderate 20
Total studies	A + B + C		240	93	19.8 (0–63) 34	57	59 55	74 69	100	80 74	Very high 78 High 9 Moderate/low 13

Abbreviations: ACC, anterior cingulate cortex; CDG, chapeau de gendarme; EZ, epileptogenic zone; FLE, frontal lobe epilepsy; FU, follow‐up; MEG, magnetoencephalography; Nb, number; NR, not reported; PNES, psychogenic non epileptic seizure; SPECT, single photon emission computerized tomography.

^a Patient seizure free but follow‐up not indicated.

FIGURE 2.

Anatomical landmarks and functional subdivisions of the ACC. (A) The cingulate cortex has historically been divided into two regions based on cytoarchitectonic Brodmann's areas (BA). The anterior cingulate cortex (ACC) comprises BA24 (orange), BA25 (blue), BA32 (green), and BA33 (yellow), while the posterior cingulate cortex includes BA23 (red), BA29 and BA30 (brown), and BA31 (dark blue). The VAC (vertical of the anterior commissure) line demarcates the boundary between the anterior and posterior regions. (B) More recent classifications divide the cingulate cortex into four functional subdivisions based on connectivity patterns. The ventral ACC (green) exhibits strong connectivity with the limbic system and is involved in emotion and mood regulation. The dorsal ACC (dark green) connects with the dorsolateral prefrontal cortex and contributes to decision‐making via the frontoparietal network. The anterior midcingulate cortex (light blue) is linked to the dorsolateral prefrontal and premotor cortices and plays a role in cognitive control and effort‐based decision‐making. The posterior midcingulate cortex (dark blue) connects to motor and somatosensory cortices and is implicated in somatosensory integration and motor preparation. The posterior cingulate cortex (brown) serves as a core hub of the default mode network and is associated with self‐referential thought and consciousness. The retrosplenial cingulate cortex (yellow), which connects to the hippocampus and parietal regions, contributes to memory encoding and body position awareness.

3.3. Patient population

Our cohort comprised 93 patients (57% males, gender not indicated in 23). Clinical characteristics are presented in Table 1. The mean age at surgery was 19.8 years (SD: 14.1, range 0–63). Among the 74 patients for whom data were available, 34 (46%) were children (≤15 years). MRI was positive in 55 patients (59%). Intracranial recordings were performed in 69 patients (74%), including SEEG in 38 (55% of invasive evaluations). All patients underwent surgery, with Engel class I outcomes achieved in 74 of them (80%). Focal lesions were identified either by imaging or histopathology in 80 patients (86%). Focal cortical dysplasia (FCD) was the most frequent diagnosis (63 patients, 68%), including FCD type 2 (FCD2) in 49/57 patients in whom the FCD subtype was provided (86% of FCD and 56% of all pathological diagnoses), followed by low‐grade glioma (n = 14), cavernoma (n = 3), and gliosis or nonspecific findings in the remaining cases.

3.4. Assessment of bias

Due to the descriptive nature of the studies, the methodology of SEEG based on anatomo‐electro‐clinical correlations, and the high rate of case reports, the overall risk of bias was high. Selection bias was rated high in 43% of cases, while assessment bias was rated high or unclear in all cases.

3.5. Confidence in EZ localization

According to the grading system, ⁴⁵ confidence in EZ localization was rated very high in 78% of cases, high in 9%, and moderate or low in 13%.

3.6. Clinical semiology

Sleep‐related epilepsy (SRE), defined as >70% of seizures occurring during sleep, was reported in 38 of 49 patients (78%) in whom this information was available. Short, repetitive seizures lasting 10–60 s (mostly 20–40) were observed in most studies reporting seizure duration (30 patients across 10 studies). Seizures lasting >1 min were reported in only 3 patients: two children recorded during status epilepticus ³⁵ or near‐continuous seizures, ²⁸ and one adult with ACC seizures spreading to the temporal lobe. In the two pediatric cases, ictal patterns varied depending on seizure frequency, and the prolonged duration was likely related to status epilepticus. Ictal semiology is detailed in Table 2 and presented in Figures 3, 4, 5.

TABLE 2.

Main ictal symptoms and signs in ACC seizures.

No. Study, N	Auras (N), type of subjective manifestation		Vocalization/verbalization/laughter	Autonomic signs	Facial expression/CDG	Automatisms motor gestural/oroalimentary	Hypermotor/complex behavior	Dystonic posturing	Head/eye deviation	Tonic–clonic/F to BTC	LOC	Post‐ictal Confusion
1–23	22	Negative emotion 12 Autonomic 18 Visual 3 Cephalic 2	17	21	14/0	21/0	NR	4	0	0	18	NR
2–19	6	Anxiety 2 Epigastric sensation 3 Deja vu 1 Unspecific 1	11	0	0	1/1	10	5	0	1	0	NR
3–10	8	Fear 2 Anxiety 3 Abdominal sensation 1 Psychic 2 cephalic 1	7–1‐5 a	9	10‐10	1	7	3	5	0	7	6
4–10	7	Fear 2 Deja vu 1 Vertigo 1 Indescribable 3	3	0	1–1	1/2	4	6	5	1	0	NR
5–5	1	Fear 1	4–0‐1	2	2	1	4	0	1	2–2	0	1
6–3	1	Fear 1	1–1	3	1	3/1	1	0	0	0–1	2	2
7–3	1	Hypogastric sensation 1	2	1	2–1	2/2	3	0	1	0	2	NR
8–3	3	Fear 3	2	3	2–2	0	3	0	0	1–1	0	0
9–2	0	NA	1–1‐1	0	1	1	1	0	0	0–1	0	0
10–2	0	NA	1	0	2–2	0	2	0	0	0	0	NR
11–1	1	Feeling of nervousness	1	1	1	0	0	0	0	0	1	0
12–1	0	NA	1–1	1	1	0	1	0	0	0	0	0
13–1	0	NA b	1	0	0	1	1	0	0	0	0	0
14–1	1	Discomfort	0	1	0	1/1	0	0	0	0	1	1
15–1	0	NA	0–0‐1 a	1	0	1	1	0	0	0	0	0
16–1	0	NA	0–0‐1	0	0	1	0	0	0	0	0	1
17–1	0	NA b	1	0	1	1	1	0	0	0	0	0
18–1	1	Cephalic	1–1	1	1	1	1	0	1	0	1	0
19–1	0	NA	1	0	0	1	1	0	0	0	0	0
20–1	0	NA	0–1‐0	0	1–1	1	0	0	0	0	0	0
21–1	1	Blurred vision	0	1	1–1	0	0	0	0	0	0	0
22–1	1	Intense fear	1	0	1–1	0	1	0	0	0	1	0
23–1	0	NA a	1	0	1	1	0	0	0	0	0	0
T 93 (%)	54 (58%)	Affective 29 autonomic 23 other 13 (45–35‐20%)	57–5‐9 (61–5‐10%)	45 (48%)	43/19 (46–20%)	40–8 (43–9%)	42/70 (60%)	18 (19%)	13 (14%)	5–5 (5%)	33 (35%)	11/36 (31%)

Note: Affective auras included fear, anguish, anxiety, nervousness, discomfort and nightmares (grouped as negative emotion in study 1 ²³ ); autonomic auras included palpitations, dyspnea, chest shake, warm sensation, shivers, goose bumps, epigastric or hypogastric sensation, heaviness in stomach.

Abbreviations: ACC, anterior cingulate cortex; CDG, chapeau de gendarme; F to BTC, focal to bilateral tonic–clonic; LOC, loss of consciousness; N, number of patients; NA, not applicable; NR, not reported; OA, oro‐alimentary; T, total.

^a Laughter was reported at the end of the seizure in these studies.

^b In these young children, an awake reaction could be associated with subjective manifestations that were not reported due to age.

FIGURE 3.

Examples of facial expressions and motor behavior changes during ACC seizures. (A) Intense fear with sitting up and screaming, (B) Grimacing face with closed eyes, chapeau de gendarme, non‐integrated hypermotor behavior; (C) Facial expression of fear, ictal pouting with open eyes, and gesticulation; (D) Hypermotor behavior with projection of the lower limbs. All pictures correspond to the earliest phase of the seizure.

FIGURE 4.

Symptom characteristics. Symptoms are listed row‐wise, with their rate of observation, confidence grade, and timing of observation visualized using color maps alongsd the actual values.

FIGURE 5.

Typicality of symptoms. The occurrence percentage of each symptom is represented by numbers and bars. Green indicates typical symptoms, meaning they occur in more than one‐third of patients, as determined by one‐sided binomial tests assessing whether the observed rate significantly exceeds 0.33. Orange indicates not meeting this threshold.

3.7. Subjective manifestations

Auras were reported in 58% of patients, with a wide range of occurrences (20%–96%) in cohorts with at least five patients, while they were not reported in 10 of 18 studies that included only one to three patients. They may be underreported in young children unable to describe them. Affective manifestations were predominant (54%) and mainly featured negative emotions, including intense fear, anxiety, nervousness, or nightmares, often associated with autonomic symptoms such as palpitations, breathing difficulties (dyspnea or shortness of breath), warmth or shivers, piloerection (goose bumps), and epigastric or hypogastric sensations. Positive emotions were rare, with only one patient describing a “smile in my belly”. Affective and autonomic subjective manifestations represented 80% of the reported auras. Other symptoms, including blurred vision, vertigo, deja‐vu, cephalic sensations, or indescribable experiences, were less frequently reported.

3.8. Objective symptomatology

Vocalizations at seizure onset were reported in 57 patients (61%), representing one of the most frequent ictal signs in ACC seizures. These included shouting, screaming, grunting, mumbling, or humming, all carrying a strong emotional component.

Verbalizations were much less frequent, occurring in five patients across five studies. ⁵ , ²⁹ , ³¹ , ³⁶ , ³⁸ They remained within the emotional field, described as “uttered loudly fearful, disgustful or swear words,” including coprolalia, calls for help, or pleas for forgiveness (e.g., “help me,” “pardon me,” and “oh my God”). An ululating pallilalia was reported in one patient. ²⁹

Laughter was reported in nine patients (10%), though only three presented with laughter at seizure onset or as a defining feature of their seizures, ⁵ , ²¹ , ²⁷ while in the six other cases, laughter was inconstant or occurred at the end of the seizure. It was described as forced and without mirth.

Autonomic signs were reported in 48% of patients but may be underestimated if not systematically assessed. The most frequent were tachycardia, pallor, flushing, suffocation, mydriasis, piloerection, and urination.

Facial expression changes (Figure 3) were reported in 46% of patients, typically appearing early in the seizure and reflecting the patient's feelings or behavior, expressing fear, threat, or aggressiveness. Notably, fear could coexist with smiling or laughter.

Chapeau de gendarme (ictal pouting) was observed in 20% of patients. This distinctive facial expression, characterized by downward‐turning lips, was also described with mimic automatisms, making it difficult to distinguish from grimacing or other oral contractions. It could be easily overlooked among the more spectacular manifestations of hypermotor seizures and may go unnoticed in nocturnal seizures, particularly when the patient is lying prone with their head on a pillow. ³⁸ As a result, its true frequency in ACC seizures may be underestimated.

Motor gestural/oral automatisms were reported in 43% of patients and were sometimes difficult to distinguish from complex motor behaviors due to overlap. These consisted of well‐coordinated, semi‐purposeful movements involving the face, lips, hands, and feet (grimacing, puckering of the mouth, lip smacking, spitting, kissing, sucking, biting, grasping, clapping, shaking hands, touching, pinching, and finger movements). These automatisms could either integrate naturally into the patient's ongoing actions and surroundings or occur independently.

Oro‐alimentary automatisms, including swallowing and chewing, were less frequent (9%).

Hypermotor (hyperkinetic) behavior was observed in 60% of patients. It was characterized by excessive movement speed and intensity, though definitions varied across studies. Some described it as sudden and spectacular agitation resembling parasomnia, ²⁴ while others did not clearly differentiate it from complex motor behaviors, leading to their combined analysis. Movements were predominantly axial and proximal, repetitive in nature, and included rocking, kicking, pedaling, pelvic thrusting, crawling, waddling, turning, and sitting up. These behaviors often expressed fear or aggression, manifesting as thrashing, escaping, running, punching, or fighting.

Dystonic posturing was observed in 19% of patients, occurring contralateral to seizure onset and never in the earliest seizure phase.

Head and eye deviation was reported in 14% of patients, also contralateral to seizure onset and occurring late in the seizure.

Tonic–clonic manifestations and focal to bilateral tonic–clonic seizures (F‐BTC) were rare (5%), occurring occasionally and only in prolonged seizures.

Loss of consciousness (LOC) was reported in 35% of patients, but its precise assessment was often challenging and not always specified. Some studies clearly identified patients who remained aware, ³² , ³⁴ , ³⁷ while others described LOC. ²¹ , ²⁵ , ³³ , ³⁸ One study indicated that patients were rarely aware throughout the entire seizure. ²⁴ LOC was sometimes inconstant, occurring in only some seizures in a given patient. Other studies highlighted cases where bizarre or strange behaviors remained integrated into the environment or reactive to outside stimuli, with or without subsequent amnesia. ⁸ , ²⁵ A striking contrast between behavioral changes and preserved awareness was emphasized in some reports.

Postictal confusion was reported in 18 studies including 36 patients, with 31% experiencing confusion. In other cases, seizures stopped abruptly without any deficit or confusion. Postictal dangerous acts were noted in a few patients, including “jabbing a box with a sharp knife while playing with his son” ²⁵ and striking at anyone attempting restraint during or after a seizure in another one. ⁵

3.9. Statistical analysis of ictal semiology

Based on the binomial test performed, vocalization/verbalization, hypermotor‐complex motor behavior, affective/autonomic auras, autonomic signs, facial expression change, motor (gestural) automatisms were present in more than a third of the patients (Figures 4 and 5). Furthermore, pairwise odds ratio comparison of symptoms (Figure 6) showed that vocalization/verbalization occurred significantly more often than 9 out of the 15 other symptoms. This was followed by hypermotor‐complex motor behavior, affective/autonomic auras, autonomic signs, facial expression change, each showing eight significant ORs.

FIGURE 6.

Pairwise odds ratios for symptom occurrence. The color scale spans from orange (odds ratio: OR <1) to green (OR >1), with white indicating values near 1. Bold labels denote statistically significant comparisons after Holm correction. Each cell value represents the OR for the symptom on the x‐axis relative to the symptom on the y‐axis.

3.10. Other findings

Some studies reported prominent interictal personality changes, including antisocial and aggressive behavior in nine adult patients, ⁵ , ⁸ , ²¹ , ³⁰ , ³⁸ leading to imprisonment in one. ²¹ Psychogenic non‐epileptic seizures coexisted with typical seizures in one patient. ³⁰ In a pediatric series, behavioral disorders were mentioned, including ADHD (attention deficit/hyperactivity disorder) in one child. ³⁴ Improvement was reported after successful epilepsy surgery and seizure control in all patients.

3.11. Anatomical and clinical correlations

Based on EZ and epileptogenic network organizations in patients with invasive recordings, as well as the chronology of ictal symptoms and signs (Table 3 and Figure 4), we postulated that affective and autonomic manifestations, vocalizations, and facial expression changes were associated with early ACC involvement. Laughter, chapeau de gendarme, motor automatisms, and hypermotor behaviors were linked to seizure onset or propagation areas. Dystonic posturing, head and eye deviation, and tonic–clonic manifestations were less frequent and observed only during propagation. LOC was reported in one‐third of patients, likely related to seizure duration and ictal discharge propagation.

TABLE 3.

Frequency and anatomo‐clinical correlations of each ictal symptom or sign.

Ictal signs or symptoms	Frequency (% of reported data, min‐max)	Overall grade level (association with ACC)	Time of occurrence (onset or during propagation)
Aura Affective/autonomic	58 (0–96) 31 (0–55)/25 (0–82)	High	Onset
Vocalization/ verbalization	61 (43–80)/5 (0–50)	High	Onset
Laughter a	10 (0–50) a	High	Onset/propagation
Autonomic signs	48 (0–91)	High	Onset
Facial expression change	46 (0–100)	High	Onset
Chapeau de gendarme	20 (0–100)	High	Onset/propagation
Motor (oral/gestural) automatisms b	43 (0–91) b	High	Onset/propagation
Hypermotor‐complex motor behavior b	60 (0–100) b	High	Onset/propagation
Dystonic posturing	19 (0–60)	Moderate	Propagation
Head‐eye deviation	14 (0–50)	Moderate	Propagation
Tonic–clonic manifestations	5 (0–40)	Low	Propagation
F to BTC	5 (0–40)	Low	Propagation
Loss of consciousness	35 (0–78)	Low	Propagation

Note: Percentages are calculated for studies including at least 2 patients; single cases are pulled in the same series.

^a In two studies, ¹⁶ , ³⁷ laughter was observed at the end of the seizure.

^b In one study, ²³ gestural automatisms and hypermotor behavior were not clearly distinguished.

4. DISCUSSION

The results of this systematic review show that affective and autonomic manifestations, vocalization, facial expression changes, complex automatisms, and hypermotor behavior with preserved awareness represent the main clinical features of ACC seizures. These symptoms or signs are encountered in about 60% of the patients included in this cohort, and their combination increases the likelihood of an ACC origin. Verbalization with emotional content, laughter without mirth, and chapeau de gendarme can also point to ACC involvement, though observed less frequently. Overall, each component of ictal semiology shows a strong emotional content, supporting the role of the mesial prefrontal regions. In contrast, dystonic, versive, and tonic–clonic manifestations are infrequent and appear secondarily during the propagation of ictal discharges. Short seizures (<60 s), often occurring in clusters during sleep, are characteristic. These findings help define the clinical picture of ACC seizures. It should be emphasized that they are consistent with the seminal descriptions, ¹ , ² , ³ , ⁴ which had already highlighted the same clinical scenarios. Furthermore, most of the selected studies included lesional cases (86%) with favorable postsurgical outcomes, reinforcing the reliability of the reported data.

The semiology markers of ACC seizures are also concordant with ACC functions, especially those of its ventral aspects, which are involved in emotions, mood regulation, and autonomic functions. ⁵ In addition, interictal behavior changes and psychiatric disturbances further shape the clinical landscape of ACC epilepsy, suggesting a distinct syndrome among focal epilepsies. However, it remains to be determined whether ACC seizures can be reliably differentiated from those originating in other parts of the prefrontal cortex (PFC).

4.1. ACC seizures as an emotional behavior pattern

Subjective and objective emotional manifestations with striking behavioral changes are predominant in ACC seizures, making them a distinctive ictal feature. However, since semiology reflects the clinical expression of seizure electrical activity, both seizure onset and propagation phases contribute to the observed signs. ⁴⁷ Therefore, ictal semiology is closely linked to the connectivity of the involved epileptogenic networks. Due to the wide afferences and efferences of the ACC, especially with the dorsolateral and orbital frontal regions, the medial temporal lobe, and the insula, distinguishing primary from secondary involvement based only on clinical semiology is challenging. In addition, network organization along an antero‐posterior gradient may lead to different clinical presentations depending on the seizure onset zone, such as seizures originating in the superior frontal sulcus spreading either to the supplementary motor area or to the ACC. ⁴⁸

Although outside the scope of this review, we examined features that could help differentiate ACC seizures from those originating in other PFC regions, including the orbito‐frontal, fronto‐polar, lateral, and medial PFC.

Orbitofrontal seizures have been characterized by sudden motion arrest, unresponsiveness, staring, loss of awareness, ipsilateral or contralateral head or head‐and‐eye deviation, and hypermotor automatisms. ¹⁵ Notably, ictal discharges arising from the orbitofrontal region may initially be clinically silent, becoming apparent only when they propagate to other structures. ⁴⁹ Fronto‐polar seizures are rarely isolated, with few descriptions available. They have been reported as “pseudo‐absences,” ² or hypermotor seizures without emotional content. ³⁷ Based on these findings, ACC seizures can be distinguished by their predominant emotional component, which characterizes both subjective and objective manifestations. In line with this view, hypermotor behavior with facial expressions of fear or anger was more frequently observed when seizure onset involved the ACC rather than other PFC areas. ³⁷ Conversely, in the brain semiology atlas database, ⁵⁰ where emotional manifestations were not included in the model, few features specifically pointed to ACC involvement.

When considering lateral versus ventromedial PFC seizures, different network‐dependent clinical expressions have been described. ¹⁷ The lateral PFC network was associated with gestural motor behavior apparently devoid of emotional content or positive emotional expression. In contrast, the ventromedial PFC, including the paralimbic temporopolar‐insular‐orbitofrontal network, presented with a fearful emotional expression and gestural motor behavior evoking a defensive or attacking reaction (fearful seizure pattern). This explosive emotional behavior has been correlated with sudden, transient desynchronization, leading to a decoupling of the orbitofrontal cortex and amygdala, effectively removing cortical control over the subcortical emotional system. ⁴⁷ These mechanisms support the key role of the frontal cortex in emotional perception and expression. Focusing on ictal emotional behavior, it has been further demonstrated that fear, anxiety, and defensive behavior were strongly correlated with involvement of the amygdala, temporal structures, and posterior orbito‐frontal cortex. ⁵¹ Moreover, fear associated with passive behavioral expression was linked to mesial temporal structures (especially amygdala), posterior orbitofrontal cortex, and/or ACC, whereas active threat responses were associated with posterior orbitofrontal cortex, ACC, dorsolateral, and ventrolateral PFC. The refinement of anatomical and electro‐clinical correlations highlights the role of network organization in seizure semiology, demonstrating the limitations of attributing clinical features to a single anatomical structure. However, hierarchy, chronology, and qualitative analysis of each sign and symptom can help localize the seizure onset more accurately.

A similar approach was used to identify a rostro‐caudal topographic gradient from the ACC to the posterior MCC, reflected in different seizure semiology. ²⁴ ACC seizures were characterized by spectacular manifestations including global body movement, goal‐directed behaviors, impaired awareness, and high emotional intensity with expressive vocalizations, closely resembling parasomnia manifestations. The authors emphasized that the networks involved in ACC seizures overlapped with those implicated in arousal parasomnias and psychiatric disorders. Interestingly, seizures predominantly occurring during sleep were observed in 78% of patients in our cohort, sustaining the role of sleep in ACC seizure expression. It can be argued that the high proportion of FCD2, which is frequently associated with SRE, may overlap with ACC involvement per se. However, FCD2 and FLE have been identified as independent risk factors for SRE, even if frontal location was predominant for FCD2, ⁵² , ⁵³ , ⁵⁴ and PFC was a common site for FCD2. ⁵⁵ Notably, in our review, the incidence of SRE (78%) was higher than that of FCD2 (56%), suggesting that sleep itself may play a specific role in triggering ACC seizures, although the dysplastic substrate may also contribute.

Moreover, psychiatric disturbances with sociopathic behaviors were reported in 12 patients, improving after successful epilepsy surgery. Another study described four patients with severe epilepsy involving the ventromedial PFC, including the ACC, who met antisocial personality disorder criteria, with behavioral improvements in those who became seizure free after surgery. ⁵⁶ According to these authors, antisocial behavior may be linked to high seizure frequency and extensive epileptogenic networks, particularly involving the mesial temporal structures (especially the amygdala) and the contralateral hemisphere.

Finally, both ictal and interictal findings highlight the high emotional component of ACC seizures, reflecting the broad involvement of limbic networks. In contrast, dystonic and tonic features were less frequent (19% and 5%, respectively) and were mostly reported in studies that included part of the aMCC. ²⁰ , ²⁴ , ³⁸ These findings support the previously described antero‐posterior gradient of ictal semiology. ²⁴

The wide range of LOC rates (0%–78%, mean 35%) raises questions and appears low compared to prefrontal seizure series, where LOC reached 88% in a recent study. ⁵⁷ This discrepancy suggests inconsistent analysis of LOC across studies, despite its importance in ictal semiology. Case reports typically provided more detailed descriptions of awareness, while cohort studies often focused on main clinical features. As a result, some studies may have missed or incompletely reported degrees of consciousness or awareness impairment. Furthermore, “impairment” is not strictly synonymous with LOC and may include subtle disturbances detected only through formal testing, such as interaction, memory, or self‐perception of awareness. Assessing consciousness in frontal lobe seizures, particularly nocturnal hypermotor seizures, therefore remains challenging even with careful observation. Finally, seizures strictly limited to the ACC are relatively rare within prefrontal regions. These limitations highlight the difficulty of reliably assessing LOC in retrospective series, particularly compared to more readily observable signs such as motor features. Nevertheless, the striking contrast between preserved awareness and complex behavior remains a characteristic feature suggestive of ACC involvement.

4.2. Anatomical considerations

Initially divided into two parts (anterior and posterior), the cingulate cortex has since been further subdivided into several regions based on cytoarchitecture and functional connectivity (Figure 2). The current “four‐region” model includes the anterior (ACC), middle (MCC), posterior (PCC), and retrosplenial cingulate cortex (RSC). ³⁹ Previous ACC landmarks were defined by the anterior vertical line, perpendicular to the anterior–posterior commissure line (VAC, Figure 2A). ³ , ¹⁶ , ⁵⁸ After the segmentation into four regions, the ACC was redefined as a more anterior structure, positioned in front of the anterior MCC (Figure 2B). ³⁹ , ⁴⁰ Functionally, it is subdivided into a ventral (affective) part, strongly connected to limbic and paralimbic regions involved in emotional processing (amygdala, hippocampus, and ventromedial PFC), and a dorsal (cognitive) part, primarily connected to dorsolateral and frontoparietal regions implicated in cognitive and sensorimotor processing. ⁵⁹ This functional dichotomy supports the distinct roles of the ACC in seizure propagation and its importance in the emergence of ictal semiology. Given the variability in anatomical and functional definitions, most studies have grouped the ACC and aMCC together. To minimize disparities between studies, we included in our review both patients with an EZ restricted to the ACC and those in whom the aMCC was also involved. This approach was supported by findings from studies that distinguished these two regions, where semiology did not significantly differ between ACC and aMCC groups. ²⁴ However, as stated before, tonic/dystonic features and preserved awareness during seizures were more frequent when the aMCC was involved, whereas emotional manifestations were less pronounced compared to pure ACC seizures. ²⁴ These differences may account for variability in semiology across studies.

4.3. Limitations

The main weaknesses in this review are due to the high selection and assessment bias. Another limitation relates to terminology interpretation, which can explain the variability in the frequency of different signs and symptoms. For example, automatisms and complex motor behaviors observed in ACC seizures are sometimes difficult to describe and have been reported inconsistently across studies. Despite efforts to establish a common glossary, certain terms, such as “automatisms”, “hypermotor” or “hyperkinetic seizure” remain poorly defined. The term “gestural motor behavior”, ¹⁷ which incorporates a descriptive framework (i.e., non‐purposeful vs. semi‐purposeful, integrated vs. non‐integrated, hyperkinetic/hypermotor movements), can help to analyze more accurately this semiology, but it was not easy to apply retrospectively in the selected studies. To account for this variability, we have acknowledged potential overlaps and grouped hypermotor and complex motor behaviors together. Another example is “ictal pouting” (“chapeau de gendarme”), a term only recently introduced into the literature. ³⁷ Previously, it was variously described as “grimace”, “mouth puckering”, “lip corner contraction” before becoming more consistently defined and reported as “typical chapeau facies” in recent sudies. ³¹ Its attribution to different frontal and extra‐frontal regions may stem from ambiguous terminology. It should be notified that in the seminal paper describing the chapeau de gendarme sign, ³⁸ the turned‐down mouth was compared to an inverted smile, emphasizing the emotional component of the facial expression. The lip corner contraction was therefore understood as a behavioral–emotional expression rather than a purely motor output driven by stereotyped pattern generators. It was also indicated that ACC could be involved primarily or secondarily, depending on the network organization which could include different part of the mesial frontal areas (from ACC proper to motor cingulate and preSMA), frontal operculum and anterior insula. This view was detailed in the recent systematic review on CDG, ⁶⁰ which supports the predominant mesial frontal involvement.

In addition, missing data such as sleep‐related and seizure duration, level of LOC, chronological order of symptoms, and post‐ictal status were not systematically assessed across all studies. These limitations may explain some discrepancies with previous reviews on the same topic. ⁶¹ However, the combination of the main clinical features reported in this review strongly supports ACC involvement with high reliability. This can be particularly crucial when EEG and imaging are non‐informative or negative.

5. CONCLUSION

ACC seizures are clinically characterized by predominant emotional, autonomic, and sudden motor behavioral manifestations with vocalizations, often featuring expressions of fear or other negative emotions. The explosive, spectacular nature of these seizures and the striking behavioral changes with preserved awareness are highly suggestive of ACC involvement. Short seizures occurring in clusters during sleep are also a characteristic feature. These key clinical features may help differentiate seizures of ACC origin from those arising in other regions of the PFC and FLE. However, due to the extensive connectivity of the ACC with limbic and paralimbic networks, distinguishing between primary and secondary ACC involvement based on clinical semiology alone remains challenging. Nevertheless, the ACC plays a pivotal role in the organization of epileptogenic networks in PFC epilepsies, making its evaluation essential in presurgical planning when invasive monitoring is indicated.

FUNDING INFORMATION

GR's research is supported by project grants from the Anna Mueller Grocholski, the Vontobel, and the Swiss National Science (SNSF: 208184) Foundations.

CONFLICT OF INTEREST STATEMENT

None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Test yourself.

1. Which peculiar ictal features suggest ACC involvement in frontal lobe epilepsy (FLE)?

• Emotional and autonomic aura

• Sudden motor behavioral manifestations

• Vocalization

• Tonic–clonic and versive manifestations

• 2Which characteristics are typical in ACC seizures?

• Long seizures (>1 mn) with post‐ictal confusion

• Clusters of short seizures

• Complex automatisms with preserved awareness

• Sleep‐related seizures

• 3How to differentiate seizures originated in ACC from other locations in FLE?

• High emotional component

• Loss of awareness

• Autonomic features

• Non‐integrated motor behavior

Answers may be found in the supporting information .

Supporting information

Data S1.

Data S2.

Chassoux F, Sallèles E, Caudron Y, Zanin A, Laurent A, Cserpan D, et al. Ictal semiology in anterior cingulate epilepsy: A systematic review. Epileptic Disord. 2025;27:883–900. 10.1002/epd2.70066

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.