Thank you Nadia, for puting these questions for open discussion.

It would be grade if we had a software for EEG, that could automatically reject artifacts and detect the critical area. Then the clinician could more reliablely accept or reject the diagnosis of epilepsy in doubtful cases. And hopefully, we will not see patients taking medication to treet only their EEG.

Best regards

Sygkliti-Henrietta Pelidou

Thank you so much for your response! So the software should hopefully be able to provide quick information about the EEG from patients under treatment (that have already been diagnosed by a traditional full review of their EEG) in orther to come up with a prompt evaluation of the effects of medication and it should also be able to support the doctor in the diagnosis of controversial cases? For what concerns artifact rejection, would you like the software to completely reject artifactual epochs or just to suppress artifacts preserving the "cleaned" epochs? Kind regards, Nadia

The software should be interactive and with a cursor placed on any particular area of the tracing, should indicate whether it was artefact or a real activity. This is because though seizure in childhood(permit my bias, I am a Paediatrician) is common but Paediatric Neurologist is rare especially in the developing countries. This can be so handy to Paediatriciand and family medicine physicians who may have to treat epilepsy before referral to the specialist.

Thank you for your suggestions Dr. Adeboye. Would it be enough to the Paediatrician to know if a pattern in the EEG is artifactual or not? How does the Paediatrician usually treat the young patients before they are evaluated by the specialist?

As noted above, the inclusion of algorithms to filter out artifacts without loosing signal is great, as is automated identification of regional generators. Since many of these pts have multiple eegs collected over time, algorithms to identify temporal changes and/or trends would be enormously informative

Thank you very much Dr Weinstein and Dr Hanafy.

Which kind of temporal changes and trends should be detected? The ictal ones or also other trends not necessarily related to seizure onset?

Can EEG epochs showing no ictal or epileptiform interictal activity carry any diagnostic information?

Thanks a lot Dr Hanafy. What do you think about the trends not necessarily related to seizure onset? Do you think that EEG epochs showing no epileptiform activity might carry any diagnostic information?

So, according to the up-to-date state of the art, the EEG in the interictal stage (far from seizure onset) is not considered informative for diagnostic purpose? In other words, no difference is supposed to be detectable between healthy subjects' EEG and patient's EEG when the patient is not going to experience a seizure in a short time? Thank you all for your inspiring replies!

Hi, sorry, but this is wrong and I cannot leave this statement on the interictal EEG as it is. Where did you get this "state of the art" information from? Maybe this is something out of context? Fact is, interictal EEG is one of the key diagnostic modalities in epilepsy and the detection of interictal epileptic discharges in EEG is highly specific! Less than 1% of healthy subjects show epileptiform discharges in EEG ( http://www.ncbi.nlm.nih.gov/pubmed/7678394 )

Hi Dr Vollmar. My sentence just aimed to check whether I properly understood what Dr Hanafy meant so, if my sentence is wrong, it's clear that I misunderstood. I absolutely didn't mean to state which is the state of the art in this field, I'm an engineer and I'm trying to learn from epileptologists. Thank you for your reply!

Dear colleagues, yes, I did indeed come into the discussion lately, but still wanted to comment on this statement before it confuses others who might also come in late. I am glad this is rectified now. EEG is a most valuable method, but also maybe the most abused diagnostic method because unfortunately many artifacts or physiological variants get reported as pathological and vice versa. This leaves many with an unfortunately low impression of the value of EEG. Hence my prompt reaction when I read "no difference between healthy subjects and patients" anywhere... Good luck with the artifact reduction!

Dear all, I'm so glad to be here exchanging ideas with you. This is such a good way to learn and adjust the aim of my research according to the point of view of who is specialized in this field and can point out which are the real problems with EEG review. If I can ask one more question: since automatic interictal spike detection was mentioned, spikes are the only informative thing we can find in interictal EEG? Best, Nadia

Thanks Dr Hanafy!

Dear Dr. Mammone,

after everybody has mentioned seizures, epilepsy, interictal epileptiform changes, I would like to mention another important EEG element: Background activity and background reactivity to exteroceptive stimulation and changes of background activity in arousals. EEG background activity has been underestimated for years, although it holds great promises especially for the diagnosis of encephalopathy. Loss of background reactivity to noxious stimulation was shown to have a high predictive value for poor outcome in patients with hypoxic-ischemic encephalopathy after cardiac arrest and resuscitation (Rossetti et al. Neurology 2009, Rossetti et al. Critical Care 2010, Rossetti et al. Annals of Neurology 2011...). However, continuous tracing and quantification of frequency of background activity and quantification or even semi quantification of background reactivity would be of great use for monitoring critically ill patients with different kinds and stages of encephalopathy.

Best regards and thank you for posting this interesting topic

R.C. Sutter

Thank you Dr Sutter for your reply! Do you think analyzing the background activity might be useful also in epilepsy?

I agree with Hanafy. Furthermore, postictal slowing might be an important abnormality indicating postictal state, although not very specific. Knowledge on grading and detection of epileptic encephalopathy is limited.

Dear all,

Let me thank you all again for your valued contribution!

Trying to summarize our discussion about what a good EEG review software should hopefully do, I would highlight the following items (not in a priority order):

- Detect critical events and discriminate between artifactual or ictal;

- Detect the critical area;

- Filter out the artifacts with negligible signal loss;

- Detect EEG Temporal trends (recurrent patterns?);

- Detect rhythmic changes in the EEG --> Follow the evolution of the rhythmic changes as regard the amplitude and frequency --> Follow the propagation of the rhythmic changes;

- Detect interictal spikes and sharp waves or spike and slow wave complexes;

- Detect post ictal slowing (as a possible lateralizing sign in ictal frontal lobe seizure).

Anything missing?

Best

Nadia