4/20/98 From Dr Michela Gallager
================================
Richard,...The Proximity measure we use for probe trials is the 
average of the 1 sec averages, that is the average distance from 
the target for the entire trial (30 sec, 60 sec whatever). 

The measure we use for the training trials is the sum of the 1 sec 
averages, that gives an approximation of the deviation from a straight 
path to the platform. So the cumulative distance is one number 
(sum of the 1 sec averages) and the proximity measure for a probe 
trial is one number (average of the 1 sec averages). Two different 
numbers. 



From Richard Baker at HVS Image:
================================
Michela I get it. The Proximity measure we provide is what you 
describe, except we compute it directly from the path, without 
using 1-sec segments. And that's to be used with probe trials. 
You are saying it's 1 measure, ie refers to the target platform-
position. OK so we do overkill & compute for all 5 possible 
platform positions.

Now we'll include also your Cumulative Distance, which will be 
1 measure. And yes we can retire those 1-second segments, and just 
sum those distance for each sample point (1/10 second). And we'll 
do 'overkill again'.

OK so...where Sum is sum of distances rat-to-platform, 
at 10 "distances" per second....

Proximity Measure = Sum / Points, ie it's an average (cms). 2003 becomes meters.

Cumulative Distance = Sum / 10, ie it's a total (cms). 2003 becomes mters.



PS at 10/2000 From Richard to all users:
========================================
So the rationale of the Gallagher Proximity Measure is that a rat might
reach the platform with moderately low latency &path length, even tho
it does not know where the platform is, just by using a sweeping search.
But the Gallagher Measure is low (good) when the rat nearly knows where
the platform is, and spends most time searching near to the platform.
So high values means less knowledge.

The Gallagher Cumulative Measure takes this further - if the rat does not
know, then it is on average further from the platform, and also takes longer
to get there.

The Gallagher Learning Index described in 1993 is none of these, but was 
computed from the Gallagher Proximity Measure by Dr Gallagher. Users
who want the rationale of this should apply to Dr Gallagher.

END