Spike detection and sorting

[Reinhard]

Hi, Everyone,

I have met some problem during test the system with our neuron culture.

1. I didn't see any spike or brust which is very confusing. Our cortical neurons came from postnatal rats and have been cultured for 3 weeks. My neurons are still healthy I think.

2. The setting of spike detection is based on the supplementary of Closed-loop, multichannel experimentation using the open-source NeuroRighter electrophysiology platform. (Threshold is 5) I placed the MEA and powered the system, then the Sorting is horaded. After that I start to record a 10min recording. Some problem happens during my analysis:

    2.1. As I mentioned, I didn't see any spike from NR, but when I load the .spk into the matlab, it shows more than 90000 spikes. Does it mean I really got so many spikes (It seems not real) or I need to do some manual sorting in matlab before or after loading? I have put the .spk in googledrive, could you have a look? (https://drive.google.com/folderview?id=0B9KXzsPytfALfmo2cUlpcU4waXJOSmRocjFfMUJhX2lldUh4Mm1TaUZlMlBobUYzVWsyalU&usp=sharing).

    2.2. The 10 min .raw data is about 1.8G. When I load it in matlab, the warning is out of memory. Is it normal? If that, should I save the data in a shorter time?

Thank you!

[Jon Newman]

o

On Fri, Mar 27, 2015 at 5:29 AM, Reinhard <laos...@gmail.com> wrote:

Hi, Everyone,

I have met some problem during test the system with our neuron culture.

1. I didn't see any spike or brust which is very confusing. Our cortical neurons came from postnatal rats and have been cultured for 3 weeks. My neurons are still healthy I think.

Looks OK, but a bit sparse for getting really strong bursting activity.

 

2. The setting of spike detection is based on the supplementary of Closed-loop, multichannel experimentation using the open-source NeuroRighter electrophysiology platform. (Threshold is 5) I placed the MEA and powered the system, then the Sorting is horaded. After that I start to record a 10min recording. Some problem happens during my analysis:

    2.1. As I mentioned, I didn't see any spike from NR, but when I load the .spk into the matlab, it shows more than 90000 spikes. Does it mean I really got so many spikes (It seems not real) or I need to do some manual sorting in matlab before or after loading? I have put the .spk in googledrive, could you have a look? (https://drive.google.com/folderview?id=0B9KXzsPytfALfmo2cUlpcU4waXJOSmRocjFfMUJhX2lldUh4Mm1TaUZlMlBobUYzVWsyalU&usp=sharing).

What exactly do you mean when you say you dont see any spikes? Do you mean that in the 'spike wfms' window you don't see anything after you have set your parameters for the spike detector?

 

    2.2. The 10 min .raw data is about 1.8G. When I load it in matlab, the warning is out of memory. Is it normal? If that, should I save the data in a shorter time?

This has nothing to do with neurorighter. You a trying to put 1.8G of raw data (basically a large matrix of floating point values) was stored on your harddrive into RAM for use in matlab. You don't seem to have enough RAM to support this. You can load data in chunks or look at single channels to get around this:

help loadraw

for usage

 

Thank you!

--
You received this message because you are subscribed to the Google Groups "NeuroRighter Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neurorighter-us...@googlegroups.com.
To post to this group, send email to neurorigh...@googlegroups.com.
Visit this group at http://groups.google.com/group/neurorighter-users.
For more options, visit https://groups.google.com/d/optout.

[Reinhard]

On Friday, March 27, 2015 at 10:12:07 PM UTC+8, Jon Newman wrote:

o

On Fri, Mar 27, 2015 at 5:29 AM, Reinhard <laos...@gmail.com> wrote:

Hi, Everyone,

I have met some problem during test the system with our neuron culture.

1. I didn't see any spike or brust which is very confusing. Our cortical neurons came from postnatal rats and have been cultured for 3 weeks. My neurons are still healthy I think.

Looks OK, but a bit sparse for getting really strong bursting activity.

 

2. The setting of spike detection is based on the supplementary of Closed-loop, multichannel experimentation using the open-source NeuroRighter electrophysiology platform. (Threshold is 5) I placed the MEA and powered the system, then the Sorting is horaded. After that I start to record a 10min recording. Some problem happens during my analysis:

    2.1. As I mentioned, I didn't see any spike from NR, but when I load the .spk into the matlab, it shows more than 90000 spikes. Does it mean I really got so many spikes (It seems not real) or I need to do some manual sorting in matlab before or after loading? I have put the .spk in googledrive, could you have a look? (https://drive.google.com/folderview?id=0B9KXzsPytfALfmo2cUlpcU4waXJOSmRocjFfMUJhX2lldUh4Mm1TaUZlMlBobUYzVWsyalU&usp=sharing).

What exactly do you mean when you say you dont see any spikes? Do you mean that in the 'spike wfms' window you don't see anything after you have set your parameters for the spike detector?

Hi，Jon, the window of spikes is just below and no spikes can be seen.  But after input the .spk into matlab, it said more than 90,000 spks were found in a 10 min record. In today's test, I tried 1 min record from amp without MEA and amp with cultured MEA. Both of them give me about 1700 spikes.

Could you give me a screenshot of spikes in the window of spikes and some raw data you don't need for me to do a analysis training?

The following steps is what I did during the test, if there's something wrong, please let me know it.

1. Start the software and then stop it.

2. Click the SALPA, and train the noise

3. Place the MEA into the amp

4. Start the software

5. Change the parameter of spike detection and sorting. 

6. Hoard the training data till more than 60 channels are sorted. (I can't sort all the channels sorted) Should I do it before I placed the MEA?

7. Train and engage the spike sorter.

8. Start to record something.

 

    2.2. The 10 min .raw data is about 1.8G. When I load it in matlab, the warning is out of memory. Is it normal? If that, should I save the data in a shorter time?

This has nothing to do with neurorighter. You a trying to put 1.8G of raw data (basically a large matrix of floating point values) was stored on your harddrive into RAM for use in matlab. You don't seem to have enough RAM to support this. You can load data in chunks or look at single channels to get around this:

help loadraw

for usage

 

Thank you!

--
You received this message because you are subscribed to the Google Groups "NeuroRighter Users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to neurorighter-users+unsub...@googlegroups.com.

[Jon Newman]

Spikes are defined as any voltage event that crosses the threshold conditions set within the spike detector form will be counted as a spike. These look like noise -- your spike detection settings are too liberal. Please refer to Table Sll of the supplementary material to understand how the spike detector works

http://journal.frontiersin.org/file/downloadfile/26827/octet-stream/Supplementary%20Material.PDF/11/5/32539

[Reinhard]

Hi, Jon,

Thank you. But the parameters I used was the typical values from this table. 

Threshold: 5

Pre-spike time: 0.5 msec

Post-spike time: 1.5 msec

Noise estimation algorithm: Fixed

Spike alignment algorithm: I didn't find the where is the setting is.

Min. spike width: 80 usec

Max. spike width: 1500 usec

Min. spike amplitude: 500 uV

Min. spike slope: 2.5 uV/samp

Dead time: 1 msec

Projection type: PCA

Projection dimension: 2

Max. K: 4

P-value: 0.01

How could I define the spike is from neuron or it's just the noise because of my liberal setting?

The most important thing is why only noise and no any signal? I reviewed the paper in JoVE, their spikes were very obvious. I never find any spike like these.

Thank you!

To unsubscribe from this group and stop receiving emails from it, send an email to neurorighter-users+unsubscribe@googlegroups.com.

[Jon Newman]

Doesn't matter. Any voltage event that meets those criteria will be counted as a spike. Again, Neurorighter (nor you, really) has no objective way to define what is truly a spike generated by a neuron and what is a threshold crossing due to a random noise fluctuation. See below:

On Thu, Apr 2, 2015 at 2:22 AM, Reinhard <laos...@gmail.com> wrote:

Hi, Jon,

Thank you. But the parameters I used was the typical values from this table. 

Threshold: 5

Pre-spike time: 0.5 msec

Post-spike time: 1.5 msec

Noise estimation algorithm: Fixed

Spike alignment algorithm: I didn't find the where is the setting is.

Min. spike width: 80 usec

Max. spike width: 1500 usec

Min. spike amplitude: 500 uV

Min. spike slope: 2.5 uV/samp

Dead time: 1 msec

Projection type: PCA

Projection dimension: 2

Max. K: 4

P-value: 0.01

How could I define the spike is from neuron or it's just the noise because of my liberal setting?

There are some papers defining what a nominal extracellular spike should look like, but honestly it depends more on the impedance characteristics of your electrode than the neuron IMO. I would say the only true signature of a 'healthy' culture would be bursting, which is extremely obvious when it occurs. Large synchronized barrages of spiking activity across most or all electrodes.

The most important thing is why only noise and no any signal? I reviewed the paper in JoVE, their spikes were very obvious. I never find any spike like these.

Its very hard to tell. It could be that your cultures are not healthy, not dense enough, or are somehow not in close proximity to the electrodes. If you have access to a patch clamping rig, it would be worth it to see if the cells are healthy using intracellular recordings.  

To unsubscribe from this group and stop receiving emails from it, send an email to neurorighter-us...@googlegroups.com.