Overcounting #R responses

[Med_Support]

WJordan
Overcounting #R responses
July 06, 2012, 04:47:16 pm

I monitor licking by rats in four boxes with ENV-251L photobeam lickometers.  I store the time of every lick in an array with a resolution of 50ms (e.g., a S.S. increments T every 50ms and the time of the lick is stored in minutes as T*.050).  I’m running MedState at the traditional 10ms temporal resolution.

Rats lick at a fairly constant rate of 7/sec, but I am recording licking rates as high as double this.  For instance, a one-sec period of licking could produce values in the array of 10.00, 10.10, 10.15, 10.15, 10.30, …).   In a 10-min session, a rat today registered 4000 licks, 800 of which had a 0.00 sec difference between consecutive values in the lick array.

I do not believe these are real values.  Has anyone else run into this when driving the MedState hard to capture rapid responses  such as these?

Thanks.

Wes Jordan
St. Mary’s College of Maryland
wpjordan@smcm.edu

[Med_Support]

Gary Bamberger
Re: Overcounting #R responses
July 10, 2012, 07:45:43 am

You are using an internal resolution of 50ms in your program which is not very accurate. The fact that you are getting responses with 0.00s in between is not surprising. If the animal licks and then licks again 30ms later it will record a time difference of 0.00s.
Are you certain that the animal is not licking 4000+ times. If the animal is very thirsty, then they are going to lick a lot.
The hardware is designed to run at a 10ms resolution and in most cases can handle a 1ms resolution. Try making contact with the lickometer with your finger or breaking the photobeam if you have that type of input. Do you get more than one input when you make contact?

[Author]

Posts