Related Experiment Video
Updated: Feb 10, 2026
05:38
Interaction between Phonological and Semantic Processes in Visual Word Recognition using Electrophysiology
Published on: June 29, 2021
2.9K
Automatic recognition of epileptic spasm via large-scale visual AI model
Jialu Xu1, Chenyu Yan1, Xiaoyan Shen1
1Department of Rehabilitation, Children's Hospital, Zhejiang University School of Medicine, National Clinical Research Center for Child and Adolescents' Health and Diseases, Hangzhou, China.
Seizure
|February 8, 2026
Summary
This study developed an AI model using Contrastive Language-Image Pre-training (CLIP) for detecting epileptic spasms (ES) in infants. The model achieved high accuracy, showing promise for early automated ES detection.
Area of Science:
- Artificial Intelligence
- Medical Imaging
- Pediatric Neurology
Background:
- Epileptic spasms (ES) are a severe form of epilepsy in infants, often challenging to diagnose.
- Early detection and intervention are crucial for improving outcomes in infants with ES.
- Current diagnostic methods for ES can be time-consuming and require specialized expertise.
Purpose of the Study:
- To develop and validate an effective visual artificial intelligence (AI) recognition model for detecting epileptic spasms (ES).
- To enhance the professionalism and convenience of ES detection through automated analysis.
- To leverage advanced AI techniques for improved diagnostic accuracy in pediatric epilepsy.
Main Methods:
- A video-centered AI model was constructed using a pre-trained Vision Transformer (ViT) with Contrastive Language-Image Pre-training (CLIP) for spatial feature extraction.
- A temporal convolution module was integrated to capture temporal dynamics, followed by a multi-layer perceptron for binary classification (normal vs. abnormal).
- The model was trained on approximately 330 hours of infant motion videos, with data splitting by infant to ensure robust validation using metrics like Precision, Recall, F-score, Accuracy, and AUROC.
Main Results:
- The CLIP-based classifier achieved high performance metrics: recall 1.00 ± 0.00, precision 0.78 ± 0.01, F-score 0.87 ± 0.01, accuracy 0.98 ± 0.01, and AUROC 0.99 ± 0.01.
- The AI model demonstrated superior performance compared to previously reported methods for ES detection.
- Case studies confirmed strong consistency between the model's predictions and expert annotations, even in complex cases with intermittent ES episodes.
Conclusions:
- An automatic motion recognition model for ES was successfully developed.
- The AI model shows significant potential for early automated detection of epileptic spasms in infants.
- This technology can contribute to more timely and accessible diagnosis of ES, improving patient care.
Related Concept Videos
01:16Vascular Spasm
3.7K
The vascular phase, also known as vasospasm, is the initial stage of hemostasis, crucial for preventing excessive bleeding when a blood vessel is injured. After a vessel is cut, nerves in the damaged area trigger pain and other sensory impulses. Simultaneously, the smooth muscles in the vessel wall contract, resulting in a vascular spasm. This contraction reduces the vessel's diameter at the injury site, slowing or stopping blood loss through the vessel wall. Vascular spasms typically last...
3.7K
01:28Automatic Processing and Automatic Social Behavior
262
Automatic processing refers to the cognitive operations that occur without conscious intent or awareness, playing a fundamental role in shaping social cognition and behavior. These processes enable individuals to navigate complex social environments efficiently by relying on mental shortcuts and pre-existing knowledge structures known as schemas. One of the most influential mechanisms underlying automatic processing is priming, which subtly activates mental representations through exposure to...
262
02:41pH Scale
80.2K
Hydronium and hydroxide ions are present both in pure water and in all aqueous solutions, and their concentrations are inversely proportional as determined by the ion product of water (Kw). The concentrations of these ions in a solution are often critical determinants of the solution’s properties and the chemical behaviors of its other solutes. Two different solutions can differ in their hydronium or hydroxide ion concentrations by a million, billion, or even trillion times. A common means of...
80.2K
01:26Scaling
596
In designing and analyzing filters, resonant circuits, or circuit analysis at large, working with standard element values like 1 ohm, 1 henry, or 1 farad can be convenient before scaling these values to more realistic figures. This approach is widely utilized by not employing realistic element values in numerous examples and problems; it simplifies mastering circuit analysis through convenient component values. The complexity of calculations is thereby reduced, with the understanding that...
596
01:22Thermometers and Temperature Scales
7.7K
Any physical property that depends consistently and reproducibly on temperature can be used as the basis of a thermometer. For example, volume increases with temperature for most substances. This property is the basis for the common alcohol thermometer and the original mercury thermometers. Other properties used to measure temperature include electrical resistance, color, and the emission of infrared radiation.
As many physical properties depend on temperature, the variety of thermometers is...
As many physical properties depend on temperature, the variety of thermometers is...
7.7K
01:22Gas Thermometers and the Kelvin Scale
6.4K
The definition of temperature in terms of molecular motion suggests that there should be a lowest possible temperature, where the average kinetic energy of molecules is zero (or the minimum allowed by quantum mechanics). Experiments confirm the existence of such a temperature, called absolute zero. An absolute temperature scale is one whose zero point is absolute zero. Such scales are convenient in science because several physical quantities, such as the volume of an ideal gas, are directly...
6.4K