When exactly does GlobalTimer activate?

RosHa

Hello Josh,

I've tried using the SetGlobalTimer/GlobalTimerTrig/GlobalTimerCancel functions to send 100Hz TTL signals from DIO ports on the Analog Input Module and/or port LED channels. With your help, it worked pretty well; however, I found that the timing of each signal varies in each trial. I checked GlobalTimer1_Start values on the SessionData, but they always indicate 0.0001 and may not be precise. Are there any ways to know when the TTL signals were sent or to control them precisely other than recording the actual outputs?

Thanks!

Josh

Hello RosHa,

What signal are you measuring as the trial-start time, to measure the relative onset time of the global timer TTLs?

I'd try the following:
-1. Add {...'BNC1', 1} to the output actions of the first state in your trial
-2. Patch state machine BNC output 1 to input Ch1 of an oscilloscope. Set oscilloscope Ch1 to DC-coupling, and 1x gain, and set the trigger to 2V on the rising edge of Ch1, in 'normal' mode.
-3. Patch the Analog Input Module's DIO output to oscilloscope Ch2, also configured as DC-coupled and 1x gain.
-4. Run the task. Once per trial you should see the latency of the AIM's DIO line with respect to trial start.

Please let me know what you find!

Thanks,

-Josh

RosHa

Hi Josh,

Sorry for the late reply. I used an Analog Input Module to collect the TTL signals and found that the signal initiation varied in each session. When 'BNC1', 1 was used, its initiation also varied, but the interval between the BNC output and the 100Hz signal was stable, so I assume the variations occurred on the recording end.

Thanks!

Josh

Hi RosHa,

Got it.

Can you please be more specific about the lack of precision you're seeing? How large are these differences on average, and how much do they vary?

I'd like to replicate your setup. Can you please send the protocol and a dataset to support@sanworks.io?
To my understanding, you have:
-A state machine connected to an analog input module v2 via Ethernet cable
-State machine BNC1 Out --> AIM Input Ch1
-AIM DIO Ch1 + GND --> AIM Input Ch2

Thanks,
Josh

Josh

Hi RosHa,

I looked at your code and data.
There's a point we should probably underscore in the docs:
https://sanworks.github.io/Bpod_Wiki/module-documentation/analog-input-module/#object-functions
A.getData() does not automatically return the previous trial's data. Instead, it returns all data captured since the last command to start logging data (A.startLogging() from the PC, or the 'L' command from the serial op menu:
https://sanworks.github.io/Bpod_Wiki/serial-interfaces/analogue-input-module-serial-interface/#state-machine-command-interface

In your protocol logging is started from MATLAB prior to each trial using A.startLogging(), so logging begins before MATLAB calls RunStateMachine(), sending a USB instruction to start the trial. To make the first sample synchronous with trial time 0 (when the device receives the start command), you need to send the 'start logging' command directly from the state machine in your first trial. In your case, you'd add ['L' 1] to your serial messages:

LoadSerialMessages('AnalogIn1',{['=' 0 0],['=' 0 1],['=' 1 0],['=' 1 1], ['L' 1]});

add {'AnalogIn1', 5} to the output actions of your first state,

sma = AddState(sma, 'Name', 'TimerTrig3', ...
    'Timer', 0,...
    'StateChangeConditions', {'Tup', 'ITI1'},...
    'OutputActions', {'AnalogIn1', 5, 'GlobalTimerTrig', '111'});

and omit A.startLogging; from your main loop.

Please let me know if this fixes the issue!

One additional thing to consider - the "logging" capability of the analog module is a legacy function that relies on the analog input module's microSD card to record analog signals during the trial, and does not capture analog signals during the inter-trial interval when the logged data is returned. A better option for many scenarios is to use USB data streaming, which will continuously stream data back to the PC via USB, with live monitoring in the scope() GUI. This example protocol demonstrates synchronized data capture with USB streaming during a simple visual 2AFC task. Unlike logging, data capture is continuous, so the '#' command is used to mark each trial start time in the analog dataset.

Best,

-Josh