Value Of High Frequency Oscillations In Pre-surgical Assessment Of Epilepsy

The central goal of surgery for the treatment of epilepsy is the removal of the epileptic focus in the brain while maintaining as much functional tissue as possible in order to maximize the quality of life for the patient. Ideally the patient will be seizure free and suffer from no serious neurologic deficit post-surgery. Surgical treatment of epilepsy generally falls into one of two categories; resective, making up the majority of epilepsy surgical procedures in which the goal is the complete resection of the focus of epilepsy in the patient's brain, and functional surgery, which is generally performed when a patient suffers from severe epilepsy but based upon preoperative assessment is not a candidate for resective surgery. Generally, a patient is considered for surgery after several attempts to control epileptic seizures through anti-epileptic drugs have failed and as a result of uncontrolled seizures the quality of life for the prospective patient is low, however for a patient to be considered a surgical candidate in pragmatic terms they must undergo extensive pre-surgical assessments. The goal of pre-surgical assessment is confirming the nature of the attacks, specifically confirming that the patient's seizures are epileptic in nature, and locating a single source for the patient's seizures (Alarcon, 2012). The recording of high frequency oscillations, a relatively new mapping technology, provides an additional pre-surgical assessment tool. Prior to surgery being considered as an option for seizure relief, the following basic criteria must be met: the patient must have a reliable diagnosis of intractable epilepsy, the seizures should be disabling and pose a direct obstruction to the patient's quality of life, the patient must have the psychological resources to handle both the pre-surgical assessment and surgery itself as determined by neuropsychiatric and neuropsychological valuation, and finally there must not be any general contradictions to neurosurgery (Alarcon, 2012). After this initial checklist is met, the clinical process that follows includes neuroimaging using structural MRI, CT, PET, and SPECT scans, interictal scalp and intracranial EEG of the patient in both awake and sleep states, ictal scalp and, when appropriate, intracranial EEG including video recording of the patient's typical seizure, magnetoencephalography, the performance of an amytal test, assessment of cortical excitability SPES, multiple modalities of functional mapping, and finally intraoperative electrocorticography. In recent years high frequency oscillations have started to become another tool in assessing both pathophysiological changes in the epileptic brain and locating epileptogenic zones in the brain. High frequency oscillations (HFOs) are used in conjunction with each of the aforementioned brain imaging techniques and measures of electrical activity to further improve the practice of pre-surgical mapping and pinpoint the seizure onset zone. HFOs are thought to reflect abnormal network-driven seizure causing activity and therefore provide a representation of the pathophysiological changes occurring in epileptic brains. Specifically, as Salami et al (2012) have shown, HFOs are, "sinusoid-like field potentials greater than 80Hz and several tens to hundreds of milliseconds in duration." HFOs are classified as ripples when oscillations are recorded between 80 and 200 Hz and fast ripples when oscillations are recorded between 250 and 500 Hz. A relatively new mapping technique compared to the standard EEG in practice today, HFOs are thought to be better markers of the seizure onset zone than interictal spikes alone (Salami et al). HFOs represent fast inhibitory postsynaptic potentials (IPSPs) from discharges of interneurons that regulate pyramidal cell firing. Specifically, they reflect fast IPSPs that occur during excitation on the soma of pyramidal cells (Bragin et al 1999). Early HFO studies began with the investigation of their potential use in pre-surgical evaluation for patients with mesial temporal lobe epilepsy. Jirsch et al (2006) used, spectral analysis and visual inspection techniques to study patients with focal epilepsy generating from the mesial temporal lobe. Jirsch et al (2006) also found that in each individual seizure (n=12) "well-localized, segmental, very high frequency band (200-500 Hz) oscillations were visually identified near the time of seizure onset from contacts in this [mesial temporal lobe] zone". In addition to this evidence of HFOs at the site of seizure onset, increased HFO activity ranging between 100-200 Hz was seen later in seizures in 75% of the patients. These results essentially set the stage for further investigation into the potential usefulness for HFOs by first displaying that HFOs could be recorded from humans with focal epilepsy using depth macroelectrodes and secondly by providing evidence that the majority of HFOs occur in regions of primary epileptogensis and rarely in regions of secondary seizure spread while a lack of high frequency activity indicates poor localization. There are several cellular mechanisms that could potentially generate a HFO. Inhibitory interneurons have the potential to pull pyramidal cells or other interneurons into a rhythmic firing pattern which could in turn activate hippocampal interneurons to produce rhythmic IPSPs and generate a HFO. Staba (2012) was able to demonstrate that blocking GABAA mediated inhibition in the hippocampus has also been shown to generate normal ripple HFOs. An additional proposed mechanism for the generation of HFOs is electronic coupling mediated through gap junctions. In vitro results show that chemical agents that have the ability to interfere with gap junction communication amongst neurons in the hippocampus have the ability to suppress ripples. Conversely, manipulating these chemical agents in order to increase conductance at the gap junction can increase HFO ripples (Staba 2012). Pathologic HFOs (pHFOs) as seen in epileptic brains differ from normal HFOs. Studies in the rat hippocampus have pointed out several key differences between the normal ripples of HFOs and the activity of pHFOs. Normal ripples tend to emerge from a relatively large area of the brain and as mentioned before the ripple reflects fast IPSPs from interneuron discharges. Noebels et al (2012) found that while pHFOs do include fast ripples containing frequencies between 250 and 600 Hz similar to HFOs, they emerge from a much more discrete cluster of neurons and during pHFOs there is no consistent firing pattern between interneurons (Staba 2012). Pathological HFOs are associated with sites of seizure onset and are linked with different types of pathology including hippocampal sclerosis, a condition of neuronal loss in the hippocampus. Pathologic HFOs reflect bursts of population spikes from local clusters of abnormally synchronized principal cells and have the potential to be used in the identification and recognition of the location of epileptogenesis. This is because pathologic HFO intensity increases during the onset of seizures and pHFOs appear early after an epileptogenic event. Both events predict the development of epilepsy. Opinions on the extent to which HFOs can aid in pre-surgical assessment are varied. Many studies involving both animal and human models suggest that HFOs are highly correlated with the epileptogenic zone, however they can be difficult to detect due to their low amplitude. Specifically, Fujiwara et al (2012) have provided evidence that the prevalence of ictal HFOs and their role in localization of the epileptogenic zone on intracranial EEG as part of pre-surgical assessment is unknown. HFOs have been in use for just over a decade, initially recorded in the hippocampus of patients with mesial temporal lobe epilepsy and since have been used to measure recordings at clinical depth with grid electrodes throughout multiple brain areas and with patients with varying types of epilepsy (Bragin et al 2010). While it is clear that HFOs can add to pre-surgical assessment, further investigation is necessary to determine how best to collect, interpret and utilize the data provided in a clinical setting. Consistent with other early explorations into seizure control protocols, most early studies with HFOs were conducted with rats treated with kainic acid. In a relatively early study, Bragin (1999) compared oscillations in kainic acid treated rats with those recorded in both kindled and normal rats. Fast ripple oscillations (ranging from 200-500 Hz) were located through correspondence with the negative phase of burst discharges of pyramidal cells only in the KA-treated rat group in areas directly adjacent to epileptogenic lesions on the hippocampus, entorhinal cortex, and dentate gyrus. The conclusion was drawn that fast ripples are actually, "field potentials reflecting hypersynchronous bursting of excitatory neurons." From this study came a suggested hypothesis for the contribution of HFOs to pre-surgical assessment. Bragin et al (2012) observed that the oscillations function to reflect fields of hypersynchronized action potentials within the neuronal clusters responsible for the generation of seizures. Therefore HFOs could be used in pre-surgical assessment as a biomarker for epileptogensis, epileptogenicity, and finally the descri ption of the epileptogenic region. This could be especially helpful in patients whose seizures are poorly localized, as is often seen in neocortical epilepsy where seizures tend to be widespread at onset. In a separate study examining patients with neocortical epilepsy, Rosenaw & Luders (2001) found that HFOs are highly localized in the seizure onset zone before the actual seizure onset. The majority of seizures in each patient (n=6) was predicted by an increase in HFO activity in the 20 minutes prior to neocortical seizure onset. HFOs may therefore be useful clinically when attempting to localize the seizure onset zone and also identify time periods where there could be increased probability of seizure likelihood. While the sample size of this particular study was too small to make any definitive statements, the results do clearly indicate that improvements can be made to current EEG systems and protocols to gain a more complete picture of both ictal and interictal activity. HFOs have also been studied in conjunction with examinations of gap junctions and their role in the generation of interictal spikes. Traub et al (2008) used three rat groups, one treated with an intracellular alkalinizing agent to open gap junctions, one with a cholinergic antagonist and a final group with the ejection of a hypertonic K+ solution. Electrical coupling between neurons through axonal gap junctions was found to underlie high frequency population oscillations in the epilepsy-prone brain. In the same experimental protocol, HFOs were again found at the site of seizure onset and noted as a, "functional indicator of the location of an epileptic focus." Furthermore this study provides evidence that HFOs aren't only useful in the location of epileptogenic zones for surgical procedures but also for the broader use of understanding cellular mechanisms to gain a better understanding of the physiologic factors that initiate a seizure. The use of HFOs in pre-surgical assessment has been shown to lead to favorable surgical outcomes in paediatric epilepsy. Fujiwara et al (2012), in a separate clinical study than mentioned before, focused on children with cortical dysplasia demonstrated that ictal HFOs that are commonly found through the use of intracranial EEG and have a localizing value aiding in guiding resection. Furthermore, there was evidence of ictal propagation immediately followed by clinical aspects of seizures suggesting that ictal HFOs have the potential to add even more valuable localizing information than interictal HFOs. Overall, the complete resection of HFOs is a positive sign for successful surgical outcome (Ibrahim 2012). The use of high frequency oscillations to localize epileptogenic zones and guide surgical procedures has the potential to increase positive surgical outcomes and should be a part of pre-surgical assessment, however further research is still necessary to determine the extent of which HFOs can be used to consistently and reliably locate the site of seizure onset. Despite accumulating evidence suggesting the positive impact that HFOs could have in the development of a surgical treatment for medically intractable cases of epilepsy, there is still widespread debate as to how effective HFOs are as biomarkers and how they can best be evaluated. Sufficient groundwork has been laid, especially with in vitro rat studies, but discovery and recording methods could still be improved in human subjects, for example to alleviate difficulties with detection. Progress regarding HFO use has been substantial over the past decade however the process of refining techniques to take this technology from promising in vitro studies in mammalian tissue to regular clinical use simply takes time. In the most broad sense, the potential for the widespread use of HFOs suggests that further advancements can be made to current diagnostic and pre-surgical assessment techniques, specifically with attempts to pinpoint epileptogenic zones with difficult to map epilepsies.