Real-Time Brain Signal Tool May Improve Epilepsy Surgery

What was studied

This study looked at whether a computer system could find very fast brain wave patterns called high-frequency oscillations, or HFOs, during epilepsy surgery. HFOs are being studied as biomarkers for the brain tissue that may be causing seizures. The goal was to see if these signals could be detected fast enough during surgery to potentially help guide what tissue is removed.

The researchers first trained the system using brain recordings taken during surgery from 22 patients in Zurich. They used a special type of computing setup called a spiking neural network on neuromorphic hardware. Then they tested it in a separate group of 54 patients in Utrecht, using a different recording device and a different anesthesia setting. They also tried the system during one live surgery to see if it could give feedback in real time.

What they found

The automated detector found HFO patterns in a way that matched fairly well with a previously used detection method in the training data. In the separate test group, postresection HFOs were found more often in patients with poor seizure outcomes than in those with good outcomes. The detector found postresection HFOs in 9 patients with poor outcomes and 1 patient with a good outcome. In one live surgery, the system was able to analyze the brain recording as it was being collected and provide HFO feedback within 1 minute after the recording ended.

Limits of the evidence

This study does not prove that using this detector during surgery improves seizure outcomes. It mainly shows that the system can detect HFOs and that postresection HFOs were associated with worse outcomes in this dataset. The training group was small, and the real-time test was reported in only one live surgery. The study also used different hospitals, devices, and anesthesia methods, which supports transferability but can also add variability. The confidence interval around the main prediction result was wide, which means there is uncertainty about how strong the association really is.

For families and caregivers

For families, this study suggests that surgeons may someday have faster tools during epilepsy surgery to help identify brain areas linked to seizures. If reliable, this could support decisions in the operating room. But this research is still an early step. More studies are needed to show that using this tool in real time actually leads to better seizure control or safer surgery.

What to watch next

Stronger evidence would come from larger studies that test whether using this real-time HFO feedback during surgery improves patient outcomes.

Terms in this summary

high-frequency oscillations (HFOs)

Very fast brain wave patterns that may be linked to the brain area causing seizures.

electrocorticography (ECoG)

A recording of brain activity taken directly from the surface of the brain during surgery.

epileptogenic zone

The area of the brain thought to be responsible for starting seizures.

resection

Surgical removal of brain tissue.

spiking neural network (SNN)

A type of computer model designed to process signals in a way that is inspired by how nerve cells communicate.

neuromorphic hardware

Computer hardware built to run brain-inspired programs efficiently and quickly.

diagnostic odds ratio

A number that shows how well a test separates people with one outcome from people with another outcome; higher numbers usually mean better separation.