0
$\begingroup$

enter image description hereI have two classes of spike, namely, excitatory spikes and inhibitory spikes. In the figure you can see their shapes.

I want to classify my spikes based on these two shapes. The shape in the figure is very clean and in our real data the shape of the spikes have some noise. I want my classification algorithm chooses the class of each spike at least with 85-90% confidence. And if it can not choose with this much confidence then calls it noise.

Does anybody have any suggestion? Thanks

$\endgroup$
  • 1
    $\begingroup$ Pls post the figure. $\endgroup$ – horaceT Sep 29 '16 at 21:24
  • $\begingroup$ Why did you put PCA and SVM tags here? $\endgroup$ – amoeba Sep 29 '16 at 21:26
  • $\begingroup$ I was wondering if I can use any of these two methods for classification. $\endgroup$ – Mina Sep 29 '16 at 21:27
  • 1
    $\begingroup$ @Mina So judging from these examples, the only difference between the two spikes is in the timing of the trough and its tail. Is that typical? If that's the case, have you thought of using the time delta from peak to trough and the half-life of the tail decay as features for a binary classification? $\endgroup$ – horaceT Sep 29 '16 at 21:50
  • 1
    $\begingroup$ +1 to @horaceT. This is a standard thing in electrophysiology, have you looked at how people do it in published papers? E.g.: researchgate.net/figure/… With the current amount of information in the question, this is almost off-topic. A useful answer would have to be based on the good domain knowledge about what features of the spike waveforms are the most reliable. If you want this to be on-topic, consider e.g. showing your real data. $\endgroup$ – amoeba Sep 29 '16 at 22:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.