Good point. I'll change the code in spm_eeg_average_tf to handle phase
separately.
Vladimir
On Sat, Dec 5, 2009 at 4:08 PM, Amir H Javadi <[log in to unmask]> wrote:
> Dear Vladimir
> OK. I've just found what caused the problem. It is my method of mean value
> calculation. Such a mistake. The link below gives a proper method of mean
> calculation for phase. Just to clarify, suppose one wants to calculate the
> mean value of say 175 and -175 degrees. Using conventional mean value
> calculation it is 0, but it is 180 degrees indeed.
> http://en.wikipedia.org/wiki/Mean_of_circular_quantities
> Have a good evening :-)
> Amir
>
>
>
> 2009/12/5 Amir Homayoun Javadi <[log in to unmask]>
>>
>> Dear Vladimir
>>
>> Thanks for your reply. Yes. I converted radian to degree for my better
>> understanding, and still I don't have a complete coverage of -180 and 180.
>> Even worth than that. It doesn't cover -90 and +90. Anything you might guess
>> that I've missed?
>>
>> I simply multiplied all the values to 180/pi
>>
>> Have a good weekend :-)
>> Amir
>>
>>
>>
>> On 4 Dec 2009, at 22:35, Vladimir Litvak <[log in to unmask]>
>> wrote:
>>
>>> Dear Amir,
>>>
>>> The phase is indeed in radians so the values should be between -pi and
>>> pi. If you average the phase using the 'vector (PLV)' option you will
>>> get values between 0 and 1. If I compute the phase on the tutorial
>>> example in SPM I get values distributed more or less uniformly between
>>> -3.14 and 3.14 as expected.
>>>
>>> Best,
>>>
>>> Vladimir
>>>
>>> On Fri, Dec 4, 2009 at 9:05 PM, Amir H Javadi <[log in to unmask]>
>>> wrote:
>>>>
>>>> Hi
>>>> Dear Vladimir :-)
>>>> Another question, this time about phase in Time-Frequency analysis. I'm
>>>> confused with the phase analysis. I assume that the result is in Radian,
>>>> right? OK. I wonder why I don't have a full range of angles from +90
>>>> degrees
>>>> to -90 degrees? In none of the channels I have 90 degrees. In some of
>>>> the
>>>> channels the min and max are even -25' and +25'. It is almost symmetric
>>>> in
>>>> most of the channels, though. Am I missing something? How come I don't
>>>> have
>>>> for example +89 degrees? The data is for -200 ms before the stimulus
>>>> onset
>>>> to 3900 ms after the stimulus onset, with 512 samples/second.
>>>> Time-Frequency
>>>> analysis with 0.2 Hz frequency steps. No digital filter except one 1 Hz
>>>> High-pass filter. Frequency range from 2-40 Hz.
>>>> Again and again thanks for your attention.
>>>> Have a good evening :-)
>>>> Amir
>>>>
>>>>
>
>
|