In recent years several new rat models of human limbic/mesial temporal lobe epilepsy have been described [1, 2, 4–7, 11, 15–17]. Unlike earlier models such as kindling in which the seizures are induced by an exogenous stimulus, these new models are characterized by seizures that occur spontaneously at random intervals. Although the spontaneity of the seizures makes these models more like human epilepsy, documentation of these seizures by direct observation is highly inefficient, and subbehavioral electrographic seizures could be missed. Continuous paper EEG and video recording have been used [5–7, 15], but these techniques are resource intensive. The slow paper speed required by long-term paper recordings limits the ability to differentiate between true seizure activity and electrical artifact. Subtle behavioral seizures are likely to be missed during rapid review of video recordings alone [16]. Ambulatory cassette EEG recordings have been used [3], but the systems require expensive proprietary hardware, and the systems have limited channels for recording (8–16). To improve the utility of the models, we developed a long-term EEG/video monitoring system to detect the electrographic seizures and document their behavioral accompaniment. The system is based on commercially available components, including a computerized EEG seizure detection system that was initially developed for human seizure monitoring [8, 9, 13]. Seizures are reliably detected and the data are reduced so that 24 h of recording can be reviewed in 30–90 min. Although the computer program is accurate, special care must be taken in system design and construction to reduce sources of electrical artifact that can cause false detections when multiple animals are recorded simultaneously on a single EEG machine. During data review it is necessary to differentiate between electrical artifact induced by animal activity from true seizure activity by key EEG patterns. Certain seizure patterns (less than 3 Hz, low amplitude) will not be detected by the seizure detection program, but the system is highly effective for typical limbic seizures and may be useful for the animal models of absence epilepsy [12, 14]. It can also be used as a continuous or intermittent EEG/physiological recording device for experiments that examine animals' spontaneous behavior and the EEG correlate (e.g. sleep/wake cycles, learning and memory tasks).