LeverReversals - version_tracker.txt
----------------------------

Version numbers are stored in:

(1) .iss file (influences setup display, Control Panel)
(2) .rc file (About box)
(3) g_strVersion
(4) strXMLConfigVersion (CONFIG VERSION ONLY)
(5) help (.hm3)
(6) web site whatsnew.shtml

Version history:

===============================================
v0.1 (10-13 Feb 2005)
===============================================

- Started (10 Feb 2005) for Wendy Chang.
- Project copied from LeverAutoshaping (so the first repository version is merely that with names changed).
- PROJECT
  Intention: caudate/Acb 6-OHA +/- intracranial drugs.
  Eventually, marm paradigm (with is-it-learned-irrelevance-or-what tests after initial reversal learning,
  i.e. introduce a third stimulus... but we'll worry about that later.)
  For now, simple operant chamber reversal task.
  
  Outcome of discussion with WC/TWR 4 Feb 2005:
  
  - trial-based (as RNC's suggestion, which will aid SDT calculation as well as the obvious advantages),
    but not necessarily just FR1 for a response, but fixed consecutive number (FCN)
    style (as TWR's suggestion). For example, require FR10 (FCN10) on one of the two levers to score a "correct"
    response.
    
  - response discrimination: L or R lever.
  
  - either simple "choose L or R", or "same or opposite side as the light" tasks.
  
  - May work up to e.g. space relevant/light irrelevant <--> space irrelevant/light relevant.
  
  - Pretraining (use, for example, the SimpleSchedules program): FR1 for 50 reinfs left, then FR1 for 50 reinfs right.
  
  - Main task:
  	ITI (houselight off)
  	TRIAL: 	houselight on
  		two levers out; light on over a lever
  		require FR1... or FR10... or FR10 all consecutive... on the correct lever
  					[i.e. option: if any responses on incorrect lever, terminate trial as incorrect.]
  	WHEN REWARDED: houselight stays on for a while; food arrives.
  	
  - Option for within-session reversal at a criterion (or simply between-session reversal).
  
  - Option for probabilistic reinforcement, as for RatBat.

... Therefore, config would look like this:

	----------------------
	Task
	 * spatial ("L [or R] is correct")
	   start with * left * right correct
	   X lights are on anyway, on random sides, as a distractor
	 * light-tracking ("Same side as light [or opposite] is correct")
	   start with * same * opposite correct

	X Flash lights, when they're being used?
	  On time (s): ___ Off time (s): ___

	X Choose correct condition (L/R) pseudorandomly, rather than randomly,
	  by drawing without replacement from a list of length 2 x ___ = [...].

	Require ___ responses on the correct lever before the trial is scored as correct.
	X Require that all responses are consecutive (i.e. no responding is allowed on the incorrect lever at all).

	X p(reward | correct) = 1; p(reward | incorrect) = 0
	  p(reward | correct) = ___; p(reward | incorrect) = ___

	X Use traylight when reward delivered (probably only sensible for pellet/dipper reinforcement!).
	When reward delivered, houselight (+/- traylight) stays on for ___ seconds or until food is collected (nosepoke in alcove).

	If a trial has not been completed within ___ seconds, terminate it as an omission (enter 0 for no time limit).

	Time between trials (s): from ___ to ___

	X Reverse within a session when ___ of the last ___ trials has been performed correctly.

	Number of trials to give: ___ (0 for no limit)    Session time limit (min): ___ (0 for no limit)

	Lever debounce time (ms): ___
	-----------------------

  ... giving this lot:

	m_strRat;
	m_iSession;
	m_strComment;
	m_strSummaryFile;
	m_strDatabase;

	m_iReinforcer;
	m_fInfusionDuration;
	m_iNumDips;
	m_fDipTime;
	m_fInterDipTime;
	m_bDipperNormallyDown;
	m_iNumPellets;
	m_iPelletPulseLength;
	m_fInterPelletTime;

	m_iTaskType;
	m_bStartWithLeftCorrect;
	m_bLightsUsedInSpatialTask;
	m_bStartWithSameCorrect;
	m_bFlashLights;
	m_fFlashOnTime_s;
	m_fFlashOffTime_s;
	m_bPseudorandom;
	m_iDWORMultiplier;
	m_iNumResponsesRequired;
	m_bResponsesMustBeConsecutive;
	m_bRewardCertainty;
	m_fPRewardIfCorrect;
	m_fPRewardIfIncorrect;
	m_bUseTraylight;
	m_fFoodCollectionTime_s;
	m_fTrialTimeLimit_s;
	m_fITIMinimum_s;
	m_fITIMaximum_s;
	m_bReverseWithinSession;
	m_iReverseCriterionX;
	m_iReverseCriterionY;
	m_iMaxTrials;
	m_fSessionTimeLimit_min;
	m_iDebounceTime_ms;

	-------------
	possible values of m_iTaskType:
									TASKTYPE_SPATIAL, TASKTYPE_LIGHTTRACKING

Done:

- main window status variables/update calls
- config dialogue (with variables, events, EnsureAppropriateThings...)
- config vars in DB
- recordset for config
- config class, with summary file, writing to DB, validation
- XML for config
- CopyConfigToDialog, CopyDialogToConfig
- work out results format
- DB fields for results
- recordset for results in DB
- individual event recording (classes, DB, etc.)
- task implementation
- help
- test

===============================================
v0.2 (21 Feb 2005)
===============================================

- Pretraining will, incidentally, be FR50 left -> FR50 right (using SimpleSchedules).
- rearrange dialogue box: can't see below "If a trial has not been completed in..."
  on an 800x600 monitor.
- facility to begin trials upon centralizing nosepoke (with traylight if selected)
  m_bNosepokeToStartTrial
- option so that FRx->reward on correct lever; FRx->punishment on incorrect lever.
  (so it's symmetric)
- summary measure: running rate (avg rate of lever pressing from first to last response on correct lever)
  One method is as a query: use
    SELECT ONLY NON-OMISSION TRIALS WHERE >1 correct/incorrect responses made
	NUMBER OF RESPONSES = CorrectResponses or IncorrectResponses, minus 1
	TIME = TrialEndTime_ms - (TrialStartTime_ms + LeverReversals_Results.LatencyFirstCorrectResponse_ms)
	RATE = number/time
  But I'll also add a LatencyToLastCorrectResponse_ms and LatencyToLastIncorrectResponses_ms.
  Then can use e.g. (LatencyToLastCorrectResponse_ms - LatencyFirstCorrectResponse_ms) as the timebase.
- ensure multiple reversal technique counts properly (e.g. if "27/30" criterion, and all trials performed correctly,
  second reversal should occur on trial 60, not trial 31). Wasn't correct, now is.

===============================================
v1.0 (8 March 2007)
===============================================

- easier compilation for users

===============================================
v2.0 (12 Jan 2009)
===============================================

- Server default changed from "loopback" to "localhost" (Windows Vista compatibility and more general standardization).

===============================================
THINGS TO DO
===============================================