Preprocessing:Stimulus

From QuB

Jump to: navigation, search
Prev: Preprocessing:Go To Point Outline Next: Preprocessing:IdlStim


The Stimulus processor prepares a signal (A/D channel) to be used in the Model as a stimulus variable. For example, if you recorded both current and voltage, and you want to connect the voltage signal to variable q in the model. The Stimulus dialog includes a preview.

The signal is idealized -- described as a succession of held constant values. Each different stimulus value gives rise to different rate constants etc, so for efficiency the goal is to describe it with as few levels as possible.

After running the Stimulus processor, link the stimulus channel to the model by:

  1. in Model Properties, Rates tab, select a rate and check the box Q (Voltage) (or P (Ligand))
  2. in Data Properties, Data tab, under Experimental conditions, enter the Stimulus channel number under Voltage,Channel.

Time-varying stimuli are understood by the newer optimizers Mac and MIP, but not by MIL.


Stimulus processing proceeds in this order:

Contents

Data selection

Stimulus channel
number of the A/D channel containing the stimulus to be idealized
Data channel
number of the A/D channel containing current; generally 0
Data source
what portion of the data file to process when you click Run.
(the preview is always of the selection)
Latency
milliseconds of delay; useful if you record stimulus output, but suspect it takes time to arrive.

Spring-dashpot filter

The input signal can drive a Spring-damper system of differential equations, as described on on Wikipedia.

Natural frequency
Damping ratio

Idealized stimulus

Clear entire idealization first
advantage: Mac runs faster because there are fewer levels
disadvantage: Stimulus has to process the whole file correctly in one go
Known amps
For simple pulses, you can give the exact values of up and down
Add amps as needed, Separated by
Other levels can be found automatically, with this precision
Min duration
Dwells shorter than this will be joined to their prior neighbor

Tension pulse model

If the input signal consists of pressure pulses, this option reshapes those pulses to approximate cell membrane tension.

"On" threshold
The signal is considered "off" below this threshold. If the threshold t is negative, then x<=t is "on."
Resting tension
"Off" signals output this level
Collapsed tension
When pressure turns off, tension falls briefly to this level
Tension / pressure
When pressure is "on" at level p, tension output is p * (tension / pressure)
Tau rising
When pressure turns on, tension rises exponentially with this time constant
Tau recovering
After pressure turns off and tension "collapses," it settles exponentially back to resting tension, with this time constant

Live Mac

With this option, macroscopic current is recomputed from the model and superimposed on the data, every time the stimulus processing changes.


Prev: Preprocessing:Go To Point Outline Next: Preprocessing:IdlStim