Prev: Ant colony algorithm (ACO)
Next: to check presence of image in a folder
From: Jesse on 19 May 2010 22:16
I've seen other people ask similar questions, but still don't understand why this isn't working.

The specific error is:
Attempted to access times(1); index out of bounds because numel(times)=0.

It occurs in a utility file from Plexon called plx2trials.

**Below is the full code:**

function data = plx2trials(TimesEvents);
% Function for parsing an Nx2 matrix of time stamps and corresponding
% event codes into a cell array format compatible with the data structure
% generated by the function ctx2mat.
%
% INPUT FORMAT:
% TimesEvents(:,1) = time stamps.
% TimesEvents(:,2) = event codes.
%
% OUTPUT FORMAT
% for each trial T,
% data{T}{1} contains spike time stamps, relative to trial onset.
% data{T}{2} contains spike event codes corresponding to those time stamps.
%
% Note that events marking the beginning and end of the trial are not
% returned, as it is assumed that they are already recovered from ctx2mat.
%
% last modified 041001
% dbtm

STARTCODE = 100;
STOPCODE = 101;

data = [];

% PARSE TE INTO A CORTEX-LIKE DATA STRUCTURE
BeginTrial = find(TimesEvents(:,2)==STARTCODE);
EndTrial = find(TimesEvents(:,2)==STOPCODE);
for t = 1:length(BeginTrial)
times = TimesEvents(BeginTrial(t):EndTrial(t),1);
codes = TimesEvents(BeginTrial(t):EndTrial(t),2);
times = times-times(1);
if length(times)>2
data{t}{1} = times(2:length(times)-1)';
data{t}{2} = codes(2:length(codes)-1)';
else
data{t}{1} = [];
data{t}{2} = [];
end
end

**End of code**

The error happens in the line : times = times-times(1);

TimesEvents is a 2489x2 double when it chokes.
And 'times' as well as 'codes' are 56x1 doubles.

First few cells of 'times':
0
10
22
26
69
70
71
76

It works just fine with other data sets. I'm really confused.
Jesse
From: dpb on 19 May 2010 22:26
Jesse wrote:
> I've seen other people ask similar questions, but still don't understand
> why this isn't working.
> The specific error is: Attempted to access times(1); index out of bounds
> because numel(times)=0.
> It occurs in a utility file from Plexon called plx2trials.
>
....

> STARTCODE = 100;
> STOPCODE = 101;
>
> data = [];
>
> % PARSE TE INTO A CORTEX-LIKE DATA STRUCTURE
> BeginTrial = find(TimesEvents(:,2)==STARTCODE);
> EndTrial = find(TimesEvents(:,2)==STOPCODE);
> for t = 1:length(BeginTrial)
> times = TimesEvents(BeginTrial(t):EndTrial(t),1);
> codes = TimesEvents(BeginTrial(t):EndTrial(t),2);
> times = times-times(1);
....

> The error happens in the line : times = times-times(1);
>
....

> It works just fine with other data sets. I'm really confused. Jesse

W/O dataset itself it's hard to confirm for absolute certain but it
would seem the problem would have to be that the find() operation didn't
retrun an EndTrial>=BeginTrial so that as the error says, there are no
elements in the times array to begin with. Hence, trying to address the
first element of a null array fails.

Try removing the semicolons from the two lines

BeginTrial = find(TimesEvents(:,2)==STARTCODE)
EndTrial = find(TimesEvents(:,2)==STOPCODE)

And rerun on a failing dataset.

An error check there might be a useful feature.

--
From: Jesse on 19 May 2010 22:37
dpb <none(a)non.net> wrote in message <ht26on$hga$1(a)news.eternal-september.org>...

> W/O dataset itself it's hard to confirm for absolute certain but it
> would seem the problem would have to be that the find() operation didn't
> retrun an EndTrial>=BeginTrial so that as the error says, there are no
> elements in the times array to begin with. Hence, trying to address the
> first element of a null array fails.


This isn't the case. The first cell in 'codes' is 100 (the start code)
and the 56th cell is 101 (the stop code).

There IS data in 'times.'

Is the problem that the first cell in 'times' is 0?
Because, logically, it Should be.
From: Jesse on 19 May 2010 23:26
"Jesse " <jsheehan(a)cnbc.cmu.edu> wrote in message <ht278f$eug$1(a)fred.mathworks.com>...
> dpb <none(a)non.net> wrote in message <ht26on$hga$1(a)news.eternal-september.org>...
>
>Try removing the semicolons from the two lines
>
>BeginTrial = find(TimesEvents(:,2)==STARTCODE)
>EndTrial = find(TimesEvents(:,2)==STOPCODE)
>

I took this suggestion, and here is the last set it displayed:

times =

470744
470744
470799
470813
470815
470818
470822
470828
470840
470990
471100
471239
471245
471273
471373
471520
471556
471683
471875
471905
471949
472011
472151
472160
472190
472373
472399
472587
472666
472800
472989
473013
473171
473179
473226
473234
473373
473440
473812
473856
473923
473992
474029
474127
474272
474338
474373
474400
474418
474457
474512
474605
474733
474768
474799
474834
474946


codes =

100
1001
1002
1001
202
1001
1001
1001
1001
1001
1002
1002
1001
1002
1002
1001
1001
1001
1002
1001
1002
1002
1002
1001
1002
1001
1002
1002
1001
1001
1002
1001
1002
1001
1001
1001
1001
1001
1001
1001
1001
1002
1002
1001
1001
1002
1002
1002
1001
1001
1002
1002
1001
1002
1001
1001
101


times =

Empty matrix: 0-by-1


codes =

Empty matrix: 0-by-1
********
Why did they both suddenly go blank?
From: dpb on 20 May 2010 09:54
Jesse wrote:
> "Jesse " <jsheehan(a)cnbc.cmu.edu> wrote in message
> <ht278f$eug$1(a)fred.mathworks.com>...
>> dpb <none(a)non.net> wrote in message
>> <ht26on$hga$1(a)news.eternal-september.org>...
>>
>> Try removing the semicolons from the two lines
>>
>> BeginTrial = find(TimesEvents(:,2)==STARTCODE)
>> EndTrial = find(TimesEvents(:,2)==STOPCODE)
>>
>
> I took this suggestion, and here is the last set it displayed:
>
> times =
>
> 470744
> 470744
....
> 474799
> 474834
> 474946
>
>
> codes =
>
> 100
> 1001
> 1002
....
> 1001
> 1001
> 101
>
>
> times =
>
> Empty matrix: 0-by-1
>
>
> codes =
>
> Empty matrix: 0-by-1
> ********
> Why did they both suddenly go blank?

Well, it isn't _exactly_ what I suggested, that's the result of the two
lines in the loop, not the two find() operations.

Here's your code snippit again...

% PARSE TE INTO A CORTEX-LIKE DATA STRUCTURE
BeginTrial = find(TimesEvents(:,2)==STARTCODE);
EndTrial = find(TimesEvents(:,2)==STOPCODE);
for t = 1:length(BeginTrial)
times = TimesEvents(BeginTrial(t):EndTrial(t),1);
codes = TimesEvents(BeginTrial(t):EndTrial(t),2);
...

It's clear the second pass fails since the first time the two arrays are
not empty.

Again, instead of looking at the times and codes arrays alone, look at

BeginTrial = find(TimesEvents(:,2)==STARTCODE)
EndTrial = find(TimesEvents(:,2)==STOPCODE)

I'm quite sure you'll find a problem there in the second instance for
this data set.

--
 |  Next  |  Last
Pages: 1 2
Prev: Ant colony algorithm (ACO)
Next: to check presence of image in a folder