Recording and analysis techniques for high-frequency oscillations

Abstract

In recent years, new recording technologies have advanced such that, at high temporal and spatial resolutions, high-frequency oscillations (HFO) can be recorded in human partial epilepsy. However, because of the deluge of multichannel data generated by these experiments, achieving the full potential of parallel neuronal recordings depends on the development of new data mining techniques to extract meaningful information relating to time, frequency and space. Here, we aim to bridge this gap by focusing on up-to-date recording techniques for measurement of HFO and new analysis tools for their quantitative assessment. In particular, we emphasize how these methods can be applied, what property might be inferred from neuronal signals, and potentially productive future directions.

Introduction

Collective neuronal oscillations of functional networks in human brain occur over a wide range of spatial and temporal scales. Neocortical networks that perform critical physiological functions are organized across spatial scales from sub-millimeter cortical columns to centimeter scale lobar structures. Extracellular microwire electrodes (∼10 to 50 μm) are widely used to record the neural activity spanning single neuron action potentials to collective oscillations of large neuronal assemblies (Buzsaki, 2004). On millisecond time scales the extracellular currents associated with single neuron action potentials are detected within a radius of ∼150 μm surrounding a micro-electrode (Buzsaki, 2004). On slower time scales (<600 Hz) the linear superposition of potentials associated with extracellular current fluctuations generate a more spatially extended local field potential (LFP). The frequency of these LFP oscillations (∼DC −600 Hz) extends far beyond the traditional electroencephalogram (EEG) (Neidermeyer and Lopes da Silva, 2011, Vanhatalo et al., 2005). Mechanisms generating the extracellular currents responsible for the LFP are varied, but primarily reflect synaptic activity (Buzsaki et al., 2003), and it is difficult to directly associate LFP characteristics (e.g. frequency, amplitude, spectrum, waveform morphology, etc.) with mechanisms, physiology, or pathology. The spatial extent of the LFP is determined by the extracellular matrix and complex geometry of extracellular current flow (Mitzdorf, 1985). The LFP can be highly localized, reflecting neuronal activity within approximately 250 μm of the recording electrode (Katzner et al., 2009) to more spatially extended (1–10 mm) activity (Kajikawa and Schroeder, 2011).

Clinical macro-electrodes placed within or on the surface of the brain record a spatial average of locally generated LFP and volume conducted activity (Kajikawa and Schroeder, 2011, Whitmer et al., 2010). Because of their relatively large surface area macro-electrodes, typically ∼1 to 10 mm², do not record single neuron action potentials or multi-unit activity. Similar to micro-electrodes, LFP oscillations detected with clinical macro-electrodes span a wide range of frequencies (∼DC −600 Hz). These clinical recordings are variably referred to as electroencephalography (EEG), intracranial electroencephalogram (iEEG), or electrocorticography (ECoG). In this review we use the term LFP to describe the wide bandwidth electrical activity and oscillations (<600 Hz) recorded directly from brain using either intracranial micro- or macroelectrodes. In keeping with current clinical terminology, we also use iEEG to describe recordings obtained with intracranial electrodes. For decades researchers largely focused on activity in the Berger bands (1–25 Hz) (Gloor, 1969, Gloor, 1975). More recent studies, however, implicate gamma frequency oscillations (gamma: 25–80 Hz) and synchrony as a fundamental mechanism of percept binding and playing a critical role in brain function (Singer, 1999, von der Malsburg, 1999) and disease (Uhlhaas and Singer, 2006). Beyond the gamma frequency range, hippocampal ripple frequency oscillations (ripple: 80–200 Hz) are believed play an important role in memory function (Buzsaki et al., 1992, Buzsaki and Draguhn, 2004, Lisman and Idiart, 1995), and somatosensory neocortex ultra-fast oscillations (>400 Hz) in sensory coding (Baker et al., 2003, Telenczuk et al., 2011).

In addition to physiological LFP it was recognized early on that human epileptic brain generates pathological interictal transients, i.e. epileptiform spikes and sharp waves, that are clearly distinct from seizures and occur without clinical symptoms (Gibbs et al., 1936, Swartz and Goldensohn, 1998, Gloor, 1969). In addition to interictal epileptiform spikes and sharp waves, wide bandwidth recordings from humans and animals with epilepsy have revealed high frequency oscillations (HFO) as a possible electrophysiological signature of epileptic brain. The initial studies of recordings from human hippocampus supported the hypothesis that HFO above 250 Hz, named fast ripple (FR) oscillations (Bragin et al., 1999a, Bragin et al., 1999b), were a unique pathological oscillation associated with epileptic brain. However, recent studies reporting physiological somatosensory evoked HFO in non-human primates, likely reflecting multiunit cortical neuronal responses (Telenczuk et al., 2011), makes the specific association of activity in the FR with pathology problematic. In addition, multiple studies have now shown that HFO in the range of physiological gamma and ripple oscillations are also increased in human epileptogenic hippocampus (Worrell et al., 2008, Crépon et al., 2010, Jacobs et al., 2010), and neocortex (Worrell et al., 2004, Jacobs et al., 2010, Blanco et al., 2011, Schevon et al., 2009).

Interictal spikes and sharp waves recorded from scalp EEG are a highly specific marker of epilepsy. For example, only 69 (0.5%) of 13,658 healthy candidates for aircrew training without a history of seizures had interictal spikes or sharp waves on a routine scalp EEG (Gregory et al., 1993). Epileptiform spikes and sharp waves are identified in scalp EEG recordings as transient waveforms produced by an abrupt change in voltage polarity occurring over a duration of less than 70 ms or 70–200 ms, respectively. While this definition is simple it is well recognized by electroencephalographers that visual detection of these is events is highly subjective. Furthermore, unlike scalp EEG were recordings from normal controls provide clear information about the specificity of interictal events, iEEG recordings are essentially limited to patients with partial epilepsy. Thus, the specificity of intracranial recorded spikes, sharp waves, and HFO as biomarkers of epileptogenic brain remains challenging.

Whether HFO recorded in epileptic brain are generated by unique pathological mechanisms, or represent aberrations of normal physiological oscillations is not clear (Le Van Quyen et al., 2006). There are currently no established criteria for distinguishing physiological from pathological HFO (Engel et al., 2009, Le Van Quyen et al., 2006). These important questions are currently active areas of research. Here we review state-of-art for electrodes, acquisition, analysis, and data mining of HFO in human partial epilepsy.