Correlation coefficient between interspike intervals

Hi,

I´m struggeling a bit calculating the correlation coefficients between the ISIs of my vector, I always get 1 with a p=1 and my basic knowledge tell me that it cannot be correct. So I had voltage values from which I extracted my ISI values in ms, now I need to calculate the correlation coefficient of the intervals.

Any help/explanation? I´m new with this so I ask for patient and understanding!

[Star Strider]

‘... my basic knowledge tell me that it cannot be correct.’

It would seem that the interspike intervals would be correlated. I would plot them as functions of the initial spike time, and see if they form a specific pattern.

It would help if you post at least a representative sample of your spike times and interspike intervals data, as well as your code, and a description of the hypothesis you are testing. We might be able to suggest a better approach than a linear correlation.

[Alba Del Rio Serrato]

Voltage = -51.9375000000000 -51.9375000000000 -51.9062000000000 -51.8125000000000 -51.7500000000000 -51.7812000000000 -51.7188000000000 -51.6250000000000 -51.6250000000000 -51.5938000000000 -51.5938000000000 -51.5000000000000 -51.5938000000000 -51.5000000000000 -51.4688000000000 -51.4375000000000 -51.5000000000000 -51.5312000000000 -51.3750000000000 -51.2500000000000 -51.3750000000000 -51.3125000000000 -51.3438000000000 -51.1875000000000 -51.3125000000000 -51.1250000000000 -51.2188000000000 -51.1250000000000 -51.1875000000000 -51.0625000000000 -51.0312000000000 -50.9375000000000 -51 -50.9375000000000 -50.9062000000000 -50.6562000000000 -50.5938000000000 -50.5312000000000 -50.5000000000000 -50.5000000000000 -50.5000000000000 -50.4375000000000 -50.5312000000000 -50.5000000000000 -50.4062000000000 -50.3750000000000 -50.4062000000000 -50.2812000000000 -50.3438000000000 -50.2188000000000

(**Firing rate =0.05 ms) So what I have done to extract the ISIs is:

dt=0.05;
 a=length(voltage)*dt; 
t=[0:dt:(a-dt)];

v_thres=-20; %set threshold 

i=0; 

j=0; 

NumSpikes=0; %hold number of spikes that have occurred 

ISI =[];

for i=2:length(voltage); 

if (voltage(i) > v_thres) && (voltage(i-1) < v_thres) 

if j==0; 

j=i; 

else NumSpikes = NumSpikes + 1; 

ISI=[ISI (i-j)*dt]; 

if ((i-j)*dt) < 1; 

disp('There is at least one value below the limit.'); 

end 

j=i; 

end 

end 

end

Now I want to calculate the correlation coefficient of intervals, I know based in what they told me about the data, that after a longer intervals seems to come a shorter one.