Introduction

Brain imaging is essential in the clinical diagnosis and therapy of status epilepticus (SE) and provides a highly useful set of tools for clinical and experimental investigations of the pathophysiology of SE. After a first episode of generalized convulsive SE (GCSE) is controlled, emergency cranial X-ray computed tomography (CT) is indicated to exclude conditions that require immediate neurosurgical intervention. Complex partial SE (CPSE) and simple partial SE (SPSE) also require structural brain imaging. Brain magnetic resonance imaging (MRI) should be performed on a nonemergency basis days or weeks after CT, to detect lesions missed on CT and to add diagnostic specificity to CT findings. Partial SE often causes focal cerebral T2 signal increases, which can be misdiagnosed as neoplasia, with the patient subjected to inappropriate surgical treatment, but which usually resolve after several weeks. Permanent brain injury due to GCSE, and possibly also to CPSE or SPSE, can be demonstrated with structural MRI and studied with a variety of imaging techniques.

The pathophysiology of SE can be studied in parallel clinical-experimental research, using brain imaging in human epilepsies and experimental models of epilepsy. In humans or animals, positron emission tomography (PET), single-photon emission computed tomography (SPECT), and functional MRI (fMRI) map regional synaptic activity levels, as reflected in glucose metabolic and blood flow alterations, which are particularly useful in mapping partial SE. The severity and distribution of some metabolic changes induced by SE can be measured in humans and animals, by performing lactate imaging with MR spectroscopy (MRS) or diffusion-weighted MRI (dwMRI), soon after GCSE is terminated. The greatest difference between imaging studies of human epilepsies and experimental epilepsies lies not in the imaging techniques themselves but rather in the ability to image untreated, progressive dysfunctions of SE and to perform baseline imaging before the onset of SE, a protocol that is ethically acceptable only in animal models. Brain imaging is useful for determining the early and late structural and metabolic sequelae of SE.

Conventional brain imaging is based on four tomographic techniques, each of which detects one type of transmitted or emitted energy to construct an anatomic map of brain structure or function: X-ray CT, nuclear magnetic resonance scanning (including structural MRI, fMRI, MRS, and dwMRI), PET, and SPECT. These techniques are noninvasive and capable of imaging the entire brain simultaneously (although MRS often samples only part of the brain, for technical reasons), with best spatial resolution ranging from approximately 1mm to 10mm. The anatomically configured images most often are viewed as a series of planar slices (tomograms), but they can also be reconstructed as three-dimensional surfaces or fields. Each small-volume element (voxel) of these brain images is displayed in a color or gray-scale intensity that represents a single value for the particular imaging modality. In general, other brain imaging techniques differ fundamentally from these four techniques, either because other techniques are invasive (e.g., autoradiography, which requires tissue destruction, and optical imaging, which requires craniotomy and incision of the dura) or because they display data that do not represent a single measured value for each voxel (e.g., volumetric dipole modeling of electrophysiologic data, which displays a set of possible solutions to the “inverse problem” of electrophysiologic signal generation). The methodological principles of CT, MRI-MRS, PET, and SPECT in brain imaging and their applications in epilepsy have been reviewed in detail elsewhere (13, 68, 93, 108).

This chapter reviews clinical studies of SE that use the four conventional brain imaging techniques. Additionally, this chapter reviews studies that use these techniques in experimental models of SE, either to elucidate the basis of SE-related neuroimaging abnormalities that occur in humans or to study the pathophysiology of SE itself. Autoradiography and other invasive techniques have demonstrated regional alterations in brain perfusion, glucose metabolism, inhibitory and excitatory neurotransmitter concentrations, and neuroreceptor availabilities during and following experimental SE, and have demonstrated reversible and irreversible neuronal injuries due to SE (62, 83, 110, 205). Brain lesions due to injuries that generate experimental SE, and the associated biochemical and microphysiologic dysfunctions, cannot be studied directly in humans. In some experimental SE models, imaging maps a specific abnormality to sites that are subsequently shown to have neuronal loss. If the same type of imaging maps the same abnormality following clinical SE, one may reasonably infer that a human likely has localized neuronal loss that cannot be directly measured during a patient's life. In particular, MRS mapping of N-acetylaspartate (NAA) density has been proposed as an in vivo surrogate for histologic neuronal densitometry. As discussed later in the chapter, transitory lactate elevations on MRS and water diffusibility decreases on dwMRI also may mark sites of permanent SE-induced brain injury.

This chapter describes currently established neuroimaging applications (mainly structural brain imaging) in the clinical care of SE, followed by a review of imaging-based pathophysiologic studies in human and experimental SE.