Seizure detection

Epilepsy is a common neurological disorder that causes seizures in people of all ages. Epilepsy rates range from about 4-10 people in 1,000 in the Western world, and up to 5 times that rate in the developing world. An epileptic seizure is a consequence of a disturbance in electrical activity in the brain, but not all seizures are caused by epilepsy. 

Research on seizures

There are many types of seizures to distinguish between; currently, researchers have identified over 40 different types. In laboratory research, rodent models of seizure activity, typically differentiate between absence seizures (a type of generalized epilepsy, in which electrical activity of the entire brain is affected), and mesial temporal lobe epilepsy (a type of partial epilepsy, which only affects one hemisphere of the brain). 

Absence seizures

Absence seizures (formerly called petit mal seizures) are mostly seen in children and typically disappear around puberty. These seizures are hardly visible and thus difficult to detect by the naked eye. Instead of typical seizure activity, patients are often thought to ‘blank out’ for a moment. To diagnose them clinically, or recognize them in laboratory research, EEG activity must be measured. 

Mesial Temporal Lobe Epilepsy 

Partial epilepsy accounts for approximately 60% of all epileptic patients, with Mesial Temporal Lobe Epilepsy being the most common form. Due to its relatively high prevalence, the disabling nature of temporal lobe seizures, and the difficulty of treating with drugs, there has been much research study on this type of epilepsy, particularly over the past 20 years.  

Study of seizures in animal research

Several types of seizures are studied preclinically to investigate origin, neurological pathways, and possible treatments. In these studies, duration, number, and the intensity of seizures are important parameters. To measure these, a combination of EEG recordings and behavioral observations are often used. 

In temporal lobe epilepsy, seizures originate from the hippocampus and the amygdala. These are easily excitable in rats and mice, either chemically or electrically. Some models, such as those for status epilepsy and febrile seizures, induce seizures in the first step, are followed by a silent period with no seizures, and then spontaneous seizures will occur in the last phase. Because the silent period varies from days to weeks, researchers often have to observe hours of video to pinpoint behaviorally when these later seizures will occur. 

Seizure scales

Many studies use the Racine scale to behaviorally classify seizure episodes in rodents. However, this scale has been modified in almost as many studies as there are publications, so while the scale has been validated over many years of research, the application and interpretation remains at the discretion of the researcher. This is why some researchers are, understandably, hesitant to use automatic detection of seizure-scales. 

Advances in animal research technologies

Still, recent advances in video tracking technologies can definitely increase the ease and efficiency of epilepsy research in several ways. EthoVision XT is an automated video tracking software program that not only tracks the animal, but also detects activity changes in the video on a pixel level. This way, even when the animal is stationary (and thus the nose point, center point, and tail base of the animal stay in the same place) activity of the body can still be detected. The software does so by comparing each video frame to the previous one detect any changes. Depending on the experimental design, this technique can be applied in several ways in seizure studies, and is also put into use to detect freezing behavior (opposite of seizure activity). 

Automatically detecting seizures

By setting certain thresholds for activity, bouts of a specific amount of activity can be automatically classified and analyzed for duration, frequency, and more. Small movements such as head nodding and bigger movements such as rearing and falling can easily be filtered out and classified separately by EthoVision XT. Thresholds are set by the experimenter, allowing him or her to fine-tune settings to the specific experiment and research question, as well as the specific species and strain at hand. 

Scanning for seizure episodes

Although activity detection can be very precise, some studies may still require detailed manual scoring of seizure behavior. In this case, the same techniques can be used to filter out bouts of possible seizures from hours of video, saving researchers the pain of having to go through all the videos manually. Manual scoring, by the researcher, at seizure events of interest can be carried out within the EthoVision XT software, but you can also use a combined set-up with The Observer XT software that offers more tools for detailed behavioral scoring. In fact, you can export videos and behavioral states from EthoVision XT into The Observer XT to easily assess data in more detail.  

Combining EEG measurements with video tracking and manual scoring

In many studies, EEG measurements are used to detect seizures, specifically the types of seizures that cannot be detected visually. Both EthoVision XT and The Observer XT are able to completely integrate physiological data, such as EEG, with behavioral scores from video tracking and activity detection. Both programs have tools to easily select, visualize, and analyze all data at once. This allows researchers, for example, to select and analyze data based on combined EEG readings and activity measurements. 

Example of set-up for seizure detection and additional behavioral tests 

In Paradiso et al. (2011), researchers used PhenoTyper cages as a home cage environment to monitor rats with induced seizures for 7 consecutive days in a row. Home cage exploration was measured by EthoVision XT, seizures were measured using EEG recordings, and Racine scale seizure severity was scored using The Observer XT. All data was combined together for analysis and visualization, providing a complete picture of seizure detection and activity over multiple days. 

References 

Paradiso, B.; Zucchini, S.; Su, T.; Bovolenta, R. Berto, E.; Marconi, P.; Marzola, A.; Navarro Mora, G.; Fabene, P.F.; Simonato, M. (2011). Localized overexpression of FGF-2 and BDNF in hippocampus reduces mossy fiber sprouting and spontaneous seizures up to 4 weeks after pilocarpine-induced status epilepticus. Epilepsia, 52(3), 572–578.