The first step for the correct characterization of the patient with LOE is the collection of information about the characteristics of the seizures, both directly from the patient and from relatives/witnesses. In particular, it is important an accurate collection of the objective and subjective manifestations of the episodes, when possible, with the support of video clips of the clinical manifestations, in order to facilitate the correct interpretation of the episodes. It is also important the identification of any risk factors for the seizures, considering those mainly typical of the adulthood/old age. In particular, clinicians should exclude the effect of alcohol abuse or withdrawal, and the presence of drug therapies that can facilitate the onset of epileptic seizures, especially benzodiazepines and neuroleptics, but also some antibiotics and other commonly used medications. Moreover, by thoroughly investigating the remote medical history of the patient, also including childhood period, it is important to exclude a past history of epilepsy.

Commonly in the elderly subjects, metabolic alterations, particularly electrolyte imbalance (19, 20), can manifest with seizures, also at high frequency or even with status epilepticus (21). These include the reduced level of the principal serum electrolytes: sodium (<115 mg/dl), calcium (<5.0 mg/dl), and magnesium (<0.8 mg/dl). Other common conditions are represented by the alteration of metabolic homeostasis: low (<36 mg/dl) or high (>450 mg/dl) serum glucose, reduced urea nitrogen concentration (<100 mg/dl) and high creatinine level (>10.0 mg/dl). Since these alterations are easily treatable, they should be promptly identified and adequately corrected.

Other conditions predisposing for the onset of seizure, also typical of the adult age, are represented by alcohol withdrawal and/or intoxication (22, 23) and the intake of illicit drugs, typically cocaine (24). When clinically suspected, the measurement of plasma alcohol levels and the search for drugs metabolites in urine or blood are important data in the diagnostic process.

In the adult-elderly age there are various conditions linked with episodic loss of consciousness. Aside seizures, other potentially life threatening diseases should be excluded (25). The main causes of non-neurological loss of consciousness, sometimes difficult to discriminate from epileptic seizures, are secondary to the alteration of the cardiovascular system. Major changes in the heart rhythm, such as atrial fibrillation and prolonged pauses, can result in lipothymia and syncope (26, 27). It is therefore essential to analyze the heart rhythm with techniques such as ECG, ECG-Holter, and tilting test.

Furthermore, chronic alterations of respiratory mechanics, in particular obstructive sleep apnea syndrome (OSAS), can predispose to the development of epileptic seizures. In fact, the chronic hypooxygenation of the brain favor the development of epilepsy (28). In subjects with OSAS, the analysis of nocturnal polysomnography with oximetry is important in order to verify the extent of sleep apnea.

Neuroradiological investigations, in particular brain Magnetic Resonance Imaging (MRI), are essential to identify any structural damage responsible for the onset of seizures (29). Although, Computed Tomography (CT) remains the neuroimaging method of choice in many countries, constant advances in MRI technology led to dramatic advances in the obtaining high quality information about the living brain, which had previously only been possible by postmortem examination (30).

Evidence-based guidelines from the American Academy of Neurology recommend immediate non-contrast CT in emergency patients presenting with a first seizure regardless age at onset (31). This is still particularly important in elderly subjects to guide appropriate acute management, especially in situations such as a hemorrhage, stroke, calcified lesions or tumors.

Even if CT scan remains of value in the absence of MRI and should be considered the neuroimaging procedure of choice under these circumstances, since mid 1980s, brain MRI is largely applied for a better characterization of the epileptic phenotype (32). In the last few years, the scientific international community proposed a standard typical protocol for epilepsy-specific structural imaging based on a balance between diagnostic accuracy and clinical feasibility (33). The main advice is that neuroimaging workup of patients with epilepsy requires a minimum set of MRI basic sequences (available on most MR scanners), regardless of the clinical setting and country. Very recently, the Neuroimaging Task Force, has emphasized the meaning of high spatial resolution and image contrast with complete brain coverage to optimally assess brain anatomy and suggested the use of a three-dimensional acquisition called the harmonized neuroimaging of epilepsy structural sequences (HARNESS-MRI protocol) (34, 35). T1 before/after contrast enhancement, T2/FLAIR and DWI sequences are useful for the identification of past or acute cerebral lesions responsible for seizures; T1-inversion recovery sequence for cortical epileptogenic malformations such as cortical dysplasia and hippocampal sclerosis. Moreover, three-dimensional (3D) sequences with isotropic voxels of 1 mm or less dramatically reduce partial volume effects, a phenomenon resulting from the presence of multiple tissue types within a given voxel. Remarkably, partial volume is unfavorable when looking for subtle cortical dysplasia, as it mimics tissue blurring, a cardinal feature of these lesions. Even the application of T2-star/GRE/SWI sequences are important for the detection of small vessel diseases, both ischemic and hemorrhagic, as Cerebral Amyloid Angiopathy (CAA) and CAA-related inflammation (36, 37). In the differential diagnosis between seizures and transient ischemic attack (TIA), it is important to carry out a study of the cerebral vessels by means of CT angiography or magnetic resonance angiography.

Furthermore, imaging and post-processing imaging helps to identify early structural alterations preceding the development of drug-resistance in mild mesial temporal lobe epilepsy with LOE and to better understand pathophysiology in such syndrome (38–41). However, it should be considered that the possibility of identifying some lesions (e.g., focal cortical dysplasia) in patients with LOE is lower compared to younger subjects (8, 42, 43).

Very recently, hybrid PET/MRI systems have the potential to combine the superior soft-tissue contrast of MRI and the metabolic features of FDG-PET in a single exam (44). This may be an ideal imaging tool for detection of the epileptic zone because it can evaluate both anatomical and functional information simultaneously. This new procedure can also be very useful to study subjects with LOE and cognitive disorders such as dementia (45).

Detection of interictal EEG epileptiform discharges (IEDs) might be crucial to support a correct diagnosis in elderly, which are more likely than younger adults to present paroxysmal phenomena that masquerade as epileptic seizures. A standard EEG recording (S-EEG) is commonly the first diagnostic step in the suspect of epileptic seizures of recent onset. Nevertheless, some data in the literature indicate a reduced S-EEG sensitivity, with a lower rate of IEDs in older epilepsy patients (46–48). Moreover, both seizure etiology and comorbid medical condition (49–51), as well as several drugs, may affect IED incidence.

Among activation procedures, an adequate duration of sleep EEG recordings seems to increase the yield of IEDs in both young and elderly epilepsy patients (49). An early assessment by 24-h home-recorded Ambulatory EEG (A-EEG) in drug-free adults with recent onset focal LOEU, showed a higher propensity in elderly patients than in younger age for spiking mainly or exclusively during deep stages of NREM sleep (48). This condition appears difficult to achieve in elderly subjects by day-time laboratory EEG recordings during spontaneous or induced sleep (47). In addition, A-EEG significantly increases the probability to detect ictal epileptic discharges (52), thus it would be opportune in elderly subjects also in order to confirm a diagnosis of epileptic seizures and to verify their possible recurrence.

Although differences in seizure features between elderly and younger adult are matter of debate (53, 54), semiology and recurrence of epileptic seizures might be more difficult to ascertain in older patients. Both physiological and psychogenic non-epileptic seizures, alone or epilepsy-associated, are probably a frequent problem in elderly. Several studies (55–59), highlighted feasibility and usefulness of Video-EEG monitoring in diagnosis and differential diagnosis of seizure disorders in elderly; nevertheless, this subpopulation appears to be underrepresented among patients referred to epilepsy centers for this type of investigation.

The analysis of cerebrospinal fluid (CSF) can have numerous applications, based on the clinical-instrumental orientation specific to the case studied. The most common application of CSF analysis relates to the suspicion of Central Nervous System (CNS) infection, such as viral or bacterial meningoencephalitis. Into the correct clinical context, the analyses focus on cell and protein count, glycorrachia, bacterioscopic, and viral examinations.

Another important application in the diagnostic process of LOE relates to the evaluation of biomarkers of amyloidosis, tauopathy and neurodegeneration. In the suspicion of a neurodegenerative disease, in particular MCI due to AD, it is important to measure the concentration of the principal biomarkers of amyloidopathy (level of amyloid-β 1-42 protein) and axonal damage (TAU and P-TAU). Moreover, if prion encephalopathy (CJD) is suspected, the assay of the protein 14-3-3 and RT-QuIC analysis in the CSF is a valid support to the diagnosis.

The analysis of CSF is also fundamental for Antibody-Mediated Autoimmune Encephalitis (AMAE) (60). Classically in these encephalopathies, patients present with rapidly progressive epilepsy, poorly responsive to antiseizure drugs (ASDs). The disease is sustained by specific autoantibodies directed against neuronal surface targets (61). The most common forms are due to antibodies against LGI1, NMDAR and GABA_B R (62). It is important to recognize AMAE and treat them promptly with immunosuppressive therapy to improve patient prognosis and outcome.

Over the last 20 years, most of neuropsychological studies were carried out in adult patients with chronic epilepsy, reporting consistent rates of cognitive impairment (63–67). The most frequent findings were memory impairment, mental slowing and attentional deficits (64, 68).

More recent studies found that cognitive deficits are very common in patients with LOE, even at time of diagnosis (69–71). The assessment of the neuropsychological profile plays a significant role in the diagnostic process of LOE, mainly in the suspicion that the onset of seizures could be the first manifestation of a neurodegenerative disease. This applies is particular to those with LOEU that, in some cases, may represent a feature of an underlying AD condition (7–9, 11, 12, 17, 72, 73).

Up to now, in clinical practice, there is no agreement regarding neuropsychological assessment, with test batteries and timing being the most crucial issues. Most epilepsy centers use different batteries, resulting in substantial heterogeneity regarding test selection and outcomes (12, 73). Measures frequently used include test of global functioning, verbal episodic memory, visuospatial memory, attention/working memory and executive functions, word fluency, visuospatial functions, and constructional praxis, and logical-perceptual abilities (7–9, 11, 17, 63–67, 74, 75).

Standardizing neuropsychological assessments for detecting cognitive impairment in LOE is a priority, since the use of different measures necessary leads to different outcomes (70, 76). The only screening tool specifically designed to detect and monitor cognitive aspects in patients with epilepsy is EpiTrack (77), a brief task including tests of attention and executive functions (i.e., response inhibition, visuomotor speed, mental flexibility, visuomotor planning, verbal fluency, and working memory). EpiTrack was originally built to evaluate long-term cognitive side effects of ASDs, and its reliability in tracking global cognitive function is still largely unexplored. To this extent, although EpiTrack represents a promising tool to monitor ASDs side effects, its use is limited as it leaves other domains unexplored (78). Such point can be critical, since deficits in LOE impact on multiple cognitive domains rather than being focused on only one (7, 9, 63, 64, 66, 73, 74, 79). A neuropsychological assessment limited to screening and/or evaluation of executive functions only can lead to an underestimation of cognitive functioning in other domains (69). Therefore, a standardized and comprehensive evaluation assessing the wide spectrum of cognitive domains (i.e., global cognition, memory, attention/executive functions, language, logical reasoning, and visuospatial skills) is needed during the diagnostic work-up of LOE patients (8).

In Table 1 we report a proposal of extensive neuropsychological battery including the most used cognitive tests available in literature, potentially applicable in all centers. To the most used tests we would suggest to add the HIV-Dementia Scale (HDS), a screening tool able to detect cognitive deficits with prevalent subcortical pattern, being complementary to MMSE in clinical practice (80). Moreover, a certain degree of discrepancy may occur in some cases between neuropsychological scores and the clinical judgment on cognitive status. To this purpose, we propose to add the Clinical Dementia Rating (CDR) scale, in order to stage clinical status.

At last, some methodological issues should be carefully taken into account in the administration of test batteries. In general, since polytherapy with ASDs may influence scores on different domains, cognitive status should be always assessed at the time of diagnosis (81). This allows to identify those who manifest cognitive impairment before starting ASDs, providing a baseline from which measure the cognitive change of the disease over time and/or the effects of subsequent treatment (66, 74, 76, 82).

Longitudinal examinations of cognitive profile are recommended to monitor clinical trajectories over time and the outcomes of ASDs. Those assessments are best be performed following an adequate interval from the last seizure or when the seizure management is stable, or before the start of ASDs. We advise against repeating neuropsychological assessment within 6–9 months, due to practice effect and potential inconsistent estimates in cognitive trajectory. This is particularly true for measures of attention, memory, speed of information processing, and higher-level executive functions (83). Serial cognitive testing with long follow-up periods, especially in patients with evidence of abnormally low CSF Aβ levels, can be useful for the stratification of risk to develop cognitive decline or dementia in this population (12, 73).

A further step in EEG application in LOE is related to advanced analysis of cortical connectivity (7). EEG network analysis consists in the modeling of brain regions and their connections into a system characterized by two main features: type of connection between regions, defined as segregation, and length of pathways, defined as integration. The more the brain regions are intertwined, the more they segregate one another. At the same time, the connections of each brain region are far from being univocal, and therefore the model takes into account how the anatomical pathways between brain regions can participate in generating EEG activity. Balancing segregation and integration EEG advanced analysis can provide a mathematical representation of brain connectivity, defined as “small world” (SW) (84–86). Since cognitive decline is associated with the loss of brain connectivity, SW analysis and EEG connectivity has been investigated as a potential tool to identify people at higher risk of dementia. EEG-related neural networks analysis with SW approach demonstrated ability to distinguish Mild Cognitive Impairment (MCI) at risk of progression to AD from the general population (84).

Recently, EEG connectivity has been investigated also in LOE. By now, only few studies investigated the role of EEG connectivity analysis in LOE patients, with particular regard to LOEU (7, 13). Preliminary evidence from small sample studies highlight that cortico-cortical network is already altered at the time of epilepsy diagnosis among those people having the highest risk of dementia. In particular, SW features of people evolving to dementia over a 5-year time were already substantially different from those preserving normal cognitive function over that time (7, 13). SW reduction in the theta and delta frequency bands could be considered a potential marker of subtle cortico-cortical disconnection, potentially identifying people with LOE at the highest risk of conversion to dementia in the years following epilepsy diagnosis (7, 13). Since SW is affordable and available in any setting, SW analysis could be proposed as potential tool to provide a stratification of the risk of dementia, potentially predicting cognitive decline among patients with LOE.

Positron Emission Tomography (PET) is useful for clinical studies in epilepsy, and specific ligands probe the pathophysiology of brain dysfunction, neurochemical and receptor abnormalities. PET with 18F-FDG has been used for over 40 years to help identifying regions of focal cerebral hypometabolism, with an output of about 60% for temporal lobe epilepsy and 30% in extratemporal lobe epilepsies. FDG-PET, measuring glucose metabolism in the brain, is commonly used in the pre-surgical process for drug-resistant focal epilepsies. Dynamic PET acquisition with a quantitative assessment, rather than static PET imaging, has been suggested to increase sensitivity in identifying areas of hypometabolism when conventional FDG-PET and MRI are unremarkable (87).

Within the second level examinations in the diagnostic process of LOE, i.e., the epilepsies with onset after age 50, FDG-PET is rarely assessed. However, we should consider that brain FDG-PET is widely used to identify specific brain areas of glucose hypometabolism in neurodegenerative processes, often associated with an epilepsy syndrome (8). Together with clinical features, including neuropsychological profile, and laboratory data, in particular CSF biomarkers of amyloidosis, tauopathy and neurodegeneration, the presence of cerebral hypometabolism leads to the diagnosis of specific dementia subtypes (88). Of note, Aβ PET can mark the presence of Aβ protein in the brain parenchyma, thus supporting the diagnosis of Alzheimer's disease in patients with LOE.

Besides clinical applications, PET with adequate ligands is a highly specialized technique actively used to clarify the pathophysiology of epilepsy and possibly identify novel treatment targets. The most active lines of research are particularly focused on the relationship between inflammation, microglial activation, and epileptogenesis. In fact, brain injury and proconvulsant events can activate microglia and astrocytes to release a number of proinflammatory mediators leading to a cascade of neuroinflammatory processes in the brain. There is evidence for the implication of glial cells in a number of molecular mechanisms of seizure precipitation and recurrence: alterations in the phenotype and function of activated microglia and astrocytes, alterations of glutamine/glutamate cycle and glutamate receptor expression, release of neuromodulatory and neuroinflammatory molecules (89). PET with translocator protein (TSPO) showed an association with microglial activation, astrogliosis and cell death. So far, imaging of neuroinflammation in the context of epilepsy has been performed in limited cases, supporting the view that activated microglia, reactive astrocytes, and inflammatory intermediates may contribute to hyperexcitability in seizure foci. Increased uptake of [11C]-(R)-PK11195, or [11C]PBR28, both TSPO tracers, was observed in regions of epileptic foci and together with focal cortical dysplasia as determined by high resolution MRI, electroencephalography recording and intraictal and interictal [18F]FDG-PET contributes in demonstrating the integrated pathological role (90, 91). These findings confirm results from invasive studies revealing a specific and persistent increase in the numerical density of activated microglia within dysplastic regions of patients with epilepsy, and support the view that the inflammatory response and proinflammatory molecules contribute to epileptogenicity of focal cortical dysplasia (92). Clinical studies with more recently developed radioligands such as [11C]- PBR28, found increased TSPO ipsilateral to seizure foci in temporal lobe epilepsy, even if further studies are needed to evaluate the clinical role of this ligand (93).

Synaptic vesicle glycoprotein 2A (SV2A) binding is reduced in brain specimen obtained by epilepsy surgery. In-vivo binding of the SV2A PET tracer [11C]UCB-J was reduced in hippocampal sclerosis, with the reduction being greater than what accounted by hippocampal atrophy and reduction in FDG uptake, suggesting a further molecular process (94). A novel PET tracer, derived from 4-[2-(phenylsulfonylamino) ethylthio]-2,6-difluoro-phenoxyacetamide labeled with 11C ([11C]K-2), was recently developed that specifically binds to AMPA receptors in vivo. Increased binding was found in the cortex of patients with mesial temporal lobe epilepsy, which correlated with the AMPA receptor protein distribution in resection specimens (95). The use of PET tracer that can visualize and quantify AMPA receptors in vivo is of potential great interest and importance in the investigation of the pathophysiology of the epilepsies and seizures.