Upgrading your E

Upgrading your E-prime system to 1.1

(compatible with win2k)

Data Glitch: Interrupted Experiments have problems saving data

Some People have lost data when their experiments were interrupted (usually by loss of connectivity between client and server). Sometimes this also happens when exiting the experiment. Here is how to avoid losing your data!!

When you see the Error Dialog Box telling you that IFIS can't save the data and asks if you want to continue exiting : DO NOT HIT THE OK button. IT'S A TRAP!
Bring up the Task Manager (CTL + ALT + DELETE ) on IFIS 1
End the e-run process. NOT THE ISM MANAGER
Your .edat file is safe and cozy in the same directory as your paradigm (.ebs)
Bow down before Keith and thank him for figuring this out.

Upgrade Instructions

0 - Please read Tony Zuccolotto's message first:

1- download the E-prime 1.1 from the PSTNET web site (follow their

instructions). Install it on your machine using their setup wizard.

They should provide you with manuals, etc.

2- If you already have IFIS, download the new IFIS file:

/net/quentin/home/hernan/data/1.0.14.14.zip from our server

(stanley.engin.umich.edu). I assume that if you have IFIS in your lab, you

have access to stanley, but if not, let me know. install it using the setup wizard.

3- Your current license numbers should work. so you should be able to

install the packages in a separate win2K machine. Note - the new and the

old IFIS cannot co-exist in the same machine! If such a machine is not

available, you'll have to uninstall your current E_prime and IFIS before the

new installation. Be careful to not delete your experiments.

4 - Copy your (.es) files to the new machine, open them in e-prime and

re-compile them. Verify that everything is running smoothly. Let me know

if things still work. They seem to run OK for the simple paradigms I've

checked, but there are also glitches ( Please see below)

5 - In order to run your scripts at the FMRI system, you will need to copy your .ebs

to a directory that has the same name as your .ebs files. For example, if your paradigm is

called "blah.es" , you'll need to put the compiled version on "ifis1" (the IFIS server) at:

C:\My Experiments\IFIS\Paradigms\Research\blah\blah.ebs

...

Thank you for your support and your patience. I think this upgrade will

make life easier for everyone.

-Luis Hernandez

PST's advice | Upgrade instructions | E-prime conversions | Back to UM fmriLab

----- BEGIN FORWARDED MESSAGE ----

(8/8/2002)

Luis,

Please find attached a diagram created in PowerPoint that is useful for

understanding the various parameters of the new IFIS programming model.

Feel free to use it in your class if you find it helpful. Some labeled

items in the diagram refer to the "TDAT file" and "TBV". The TDAT file

is a new file that is written which is can be used by "Turbo

BrainVoyager (TBV) to allow real-time analysis on some properly equipt

and supported scanners. Turbo BrainVoyager is still in beta testing

but MRI Devices hopes to make it available for sale or upgrade sometime

in Q3 of this year.

When you install the new system you will notice that there are actually

two different Package File distributed and installed (IFIS & IFIS1).

Legacy paradigms will use the IFIS Package (IFIS.epk) which is mainly

designed to allow old paradigms to continue to run with the new "IFIS

System Manager (ISM)" software. The IFIS System Manager is a new

client/server application that completely replaces the old "IFIS Control

Panel (ICP)" utility. The ICP utility is no longer maintained or

supported. Legacy paradigms made need to be modified in some areas to

achieve full compatibility. PST can help resolve these problems.

New paradigms should use the IFIS1 Package (IFIS1.epk) which is designed

to take full advantage of the new paradigm programming model. This is

the only package file that will move forward in the future with respect

to maintenance and rolling out new features.

The "old IFIS" paradigm model defined paradigms in terms of Sessions,

Tasks, Blocks, Periods, and Events. Timing was mainly controlled by

the parameters of NumberOfPeriods, PeriodADuration and PeriodBDuration.

While that model was sufficient at the time, it was not a flexible model

for PST/MRIDC to build upon in the future.

The "new IFIS1" paradigm model defines paradigms in terms of Sessions,

Runs, Blocks, and Trials (as well as SubTrials, SubSubTrials if

additional levels happen to be required). There is an "Event" level

also available in the model, but it is now considered a "meta-level",

e.g. for event related paradigms the user can define any sub-section of

time within the stimulus presentation sequence to represent the "event

of interest".

By far the best and easiest way to create paradigms under the new IFIS1

model is to use the new IFIS Paradigm Wizard (covered by Tutorial 1a &

Tutorial 2).

There is also a new IFIS Import Wizard that can be used to convert

legacy paradigms to the new model, but at this time the documentation on

that process is not complete. The general approach is to actually

delete all the old model IFIS related objects and list attributes from

the paradigm, run it through the IFIS Import Wizard, and then manually

move the objects into place. Without specific documentation to guide

you the first time through this process it would be very difficult to

complete the task correctly. If you generated a paradigm using the

IFIS Paradigm Wizard first and then examined where objects need to be

placed you might be able to get through it, but my best advice here is

to pretty much ignore the IFIS Import Wizard for now until Tutorial 1c

is completed and distributed.

One other caution...

I have released these materials assuming they are primarily going to be

used for your class. I would not immediately install/upgrade your IFIS

system at the MR Center until you have users verify that their currently

running paradigms function under the current system. The best way to

do this is to designate a desktop machine in the lab as a test machine.

Everyone then goes to that machine to work with their old paradigms

until they run successfully. After all critical paradigms are running

with the new system then you should arrange a time to upgrade the system

at the MR center. Unfortunately there is no possibly way for the old

system software and the new system software to co-exist.

Hope some of this helps. If not please feel free to get us involved as

needed.

Thanks,

Tony

(back to the top)

PST's advice | Upgrade instructions | E-prime conversions | Back to UM fmriLab

Notes on E-prime conversion glitches

0 - check out this power point file to see the difference in the new

way of creating paradigms. Things have been shuffled around a little bit.

1 - You can't have both IFIS and IFIS1 packages in the same script. They do

not play well together.

2 - The Error "unrecognized BRU configuration."

After the conversion to new format, the IFISInit package call contains a bad

argument line:

c, "2, 10000, 10000, scanner,bru, on, 1.0, 1000, none, duplicate,yes"

has too many parameters. It should read:

c, "2, 10000, 10000, keyboard, bru, on, 1.0, 1000, duplicate, yes"

3 - The IFISinit (at the very beginning of the e-prime program) won't work

in the new version. If you delete the old one and then put in a new one,

that helps.

4 - If you have any objects in your program that are named after words that

are important to IFIS, the program will crash. For example, I had a

fixation mark as a text object in my program that was called "fix".

Apparently "fix" means something to IFIS, so you can't use that as a name

for something. "instr" is also a problem (and I had named my instructions

that). It's not obvious from the error message that that's the problem.

5 - You cannot run E-prime on IFIS2 (the client). If you need to open an e-prime

file while at the IFIS system, use IFIS1 (the server).

6 - Your data will be saved on IFIS2

7 - If a program crashes and you need to restart the paradigm, we find that it's better to

kill and restart the ISM daemons on both IFIS1 and IFIS2

From Jim Kroger (Psychology dept.) 12/12/02

Further beta and 1.1 incompatibilities/improvements (just Eprime not IFIS):

A big benefit of moving to new Eprime is that in the beta, even rather small paradigms were "too big to compile" which is supposedly fixed as long as you select "embedded" in the properties box for each list.

You can no longer use comma or other punctuation marks as response keys in 1.1

If you use space bar as an allowable or correct response, when 1.1 upgrades the script, it will replace the ' ' with '{SPACE}' (replaces space with {SPACE}) but if you have space in a group of responses (e.g., 'ab d' or 'abc ') this does not work, and you must by hand fix every place you had ' ' in your properties throughout the script (replace it with {SPACE}).

You can not use the word "trial" in a script, it is now reserved. You should change it in the beta version, because if 1.1 upgrades a script it omits every instance of whatever is named 'trial' and everything below it in a hierarchy.

In any slide, whereas beta would allow you to accidentally call two text boxes in the same slide by the same name (e.g., "text 4" and "text 4") without complaining, it will now crash 1.1.

In 1.1, if in a list the stimuli in a column (attribute) are lower case and the correct or allowable letters for a response are upper case (or vice versa) the program will crash, whereas in beta this was permitted.

In beta if you tried to copy and paste a text box, image, etc. from a slide to the same or another slide, it would crash Eprime. In 1.1, it works.

PST's advice | Upgrade instructions | E-prime conversions | Back to UM fmriLab