Return to the Visual Statistical Learning page
						
								VISUAL STATISTICAL LEARNING II
SCRIPT INFO

Script Author: Katja Borchert, Ph.D. (katjab@millisecond.com) for Millisecond Software, LLC
last updated: 08-15-2016 by K.Borchert (katjab@millisecond.com) for Millisecond Software LLC

Copyright © 08-15-2016 Millisecond Software

Millisecond Software thanks Turk-Browne et al. for providing the original stimuli!
 
BACKGROUND INFO

											*Purpose*
The script "Visual Statistical Learning II"  implements the procedure used by Turk-Browne et al (2005) to study the role of attention 
on the unprompted statistical learning of temporal relationships inherent in a visual presentation of shapes. 
Turk-Browne et al (2005) expand on research done by Fiser and Aslin (2002) 
(-> see Inquisit script "Visual Statistical Learning" that implements the procedure used by Fiser & Aslin).

For a more detailed description:
Turk-Browne et al (2005). The automaticity of visual statistical learning. Journal of Experimental Psychology: General, 134, 552-564.


											  *Task*
Participants are shown a sequence of 624 shapes, one shape at a time (at a presentation rate of 400ms each with a posttrialduration of
200ms), that are either green or red. The participant's task is to 
1. attend to only one of the colored shapes (green or red, counterbalanced) and 
2. press the Spacebar whenever a shape in the attended color repeats itself, regardless of how many shapes of 
the other color may have been inbetween the repeating shapes.
The order of the shapes is carefully assembled and contains a repeated number of 4 shape triplets in green and red.

In the Forced-Choice part of the experiment, the script pins the 4 familiar triplet pairs (ABC, DEF, GHI, JKL - from the sequence) 
for the attended and unattended color against 4 novel triplets (AEI, DHL, GKC, JBF, see Turk-Browne et al). 
These novel triplets are made up of shapes that have never directly followed each other during the sequence presentation. 
All shapes in the Forced-Choice task are presented in black.	
	

DATA FILE INFORMATION: 
The default data stored in the data files are:

(1) Raw data file: 'VSL_TurkBrowne_raw*.iqdat' (a separate file for each participant)

build:							Inquisit build
computer.platform:				the platform the script was run on
date, time, subject, group:		date and time script was run with the current subject/groupnumber 
blockcode, blocknum:			the name and number of the current block
trialcode, trialnum: 			the name and number of the currently recorded trial
									(Note: not all trials that are run might record data; by default data is collected unless /recorddata = false is set for a particular trial/block) 
values.attendstream:			which letter sequence is the attended one (1 or 2)
values.colorattend:				which color is the attended one (1 = green, 2 = red)
blockcode, trialcode			the name of the current block/trial
values.selectedletter:			the letter selected in PART I
stimulusitem:					the first presented stimulus in PART I and PART II
values.target_active:			whether a target was active in PART I
values.rt_target:				accumulated latencies (in ms) since the last target presentation in PART I (resets to 0 with each target presentation)
values.pairing:					which 6 shapes are presented/compared  during PART II (and in which order)																
values.shape1 - values.shape6	the six shapes in PART II
latency:						the response time (in ms)
response:						the response made by participant
correct:						the correctness of response (PART II)
values.Hit:						whether the response is counted as a Hit in PART I
values.FA:						whether the response is counted as a False Alarm in PART I

(2) Summary data file: 'VSL_TurkBrowne_summary*.iqdat' (a separate file for each participant)

script.startdate:				date script was run
script.starttime:				time script was started
script.subjectid:				subject id number
script.groupid:					group id number
script.elapsedtime:				time it took to run script (in ms)
computer.platform:				the platform the script was run on
/completed:						0 = script was not completed (prematurely aborted); 1 = script was completed (all conditions run)
/hitrate:						hit rate for detecting targets in the stream when a target was presented within the last 3s
/meanRT_hit:					mean hit latency (in ms) for detecting targets in the stream when a target was presented within the last 3s
/SD_hit:						standard deviation of hit latencies
/propcorrect_attend:			proportion correct of forced choice trials for triplets in attended stream
/propcorrect_unattend:			proportion correct of forced choice trials for triplets in unattended stream


EXPERIMENTAL SET-UP
PART1 & PART 2:
24 shapes (originals from Turk-Browne et al)* are divided into 12 green and 12 red shapes (assignment is randomly done for each participant). 
The 12 shapes in each color are then divided into 4 triplets: ABC, DEF, GHI, JKL

PART I: Shape Sequence 
There are 2 individual shape sequences of 312 items that 
1. get assigned to either the red shapes or the green shapes (counterbalanced) and
2. get interweaved randomly  (with one constraint: remaining streams shouldn't differ by more than 6 shapes) for each participant. 
The result is the observed stream of 624 shapes. Each shape is presented for 400ms with a Stimulus Onset Asynchrony (SOA) of 600ms.

Each of those 2 individual sequences were constructed such that
1. all triplets are presented 24 times
2. 6 of those 24 triplet repetitions double the last letter (e.g. ABCC, DEFF etc) -> 6 x 4 (# of triplets) = 24 extra letters, 
3. no immediate repeats of triplets (e.g. ABCABC)
4. no  repeat of triplet pairs (e.g. ABCDEFABCDEF)
(24 (#repetitions) x 4 (#of triplets) x 3 (#of letters/triplet) + 24 extra letters = 312 letters in a sequence)

Scoring of Responses:
If the participant hits the spacebar within 3000ms (default, see Turk-Browne et al) of the onset of a target (=repeated shape in the 
attended color), the response is scored as a Hit. Any other responses outside of that responsewindow are scored as a False Alarm.
Should a new target appear during the 3000ms responsewindow (and the participant has not responded to the first target), 
the responsewindow is updated for the new target and a possible response counts towards the new target.

PART II: Forced-Choice
In the Forced-Choice part of the experiment, the script pins the 4 familiar triplet pairs (ABC, DEF, GHI, JKL - from the sequence)
for the attended and unattended color against 4 novel triplets (AEI, DHL, GKC, JBF, see Turk-Browne et al). 
These novel triplets are made up of shapes that had never directly followed each other
during the sequence presentation. All shapes in the Forced-Choice task are presented in black.

Participants get to view 2 sequences of 3 shapes each. Each shape is presented for default=400ms  in the middle of the screen with a 
break of 1s  inbetween sequences. Participants are asked to decide which one was the more familiar one 
(default TASK: press "A" for first, press "L" for second -> change keys under values).
Each of the 16 triplet-pairings for the attended and unattended color is presented twice, 
once with the familiar sequence being presented first and once with the familiar 
sequence being presented second. The resulting 64 possible forced-choice tasks are selected at random.


STIMULI
Millisecond Software thanks Turk-Browne et al. for providing the original 24 stimuli in red, green, and black!
see section Editable Stimuli

INSTRUCTIONS
see section Editable Instructions

EDITABLE CODE:
check below for (relatively) easily editable parameters, stimuli, instructions etc. 
Keep in mind that you can use this script as a template and therefore always "mess" with the entire code to further customize your experiment.

The parameters you can change are:

SOA_shapepresentation:			Stimulus Onset Asynchrony (SOA; in ms) in Part I (default: 600ms)
										Note: needs to be at least 400ms (=shape presentation time)
waitforcedchoice:				determines the pretrialpause (in ms) before the first shape triplets are shown in the forced-choice task (default is 1s)
leftkey:						the scancode of the left key used to respond if the first sequence is more familiar (default = 30 -> letter "a")
rightkey:						the scancode of the right key used to respond if the second sequence is more familiar (default = 38 -> letter "l")
responsewindow:					determines how long (in ms) after a target is presented a response is counted as a Hit (default: 3000ms)


Copyright © Millisecond Software. All rights reserved.
Contact | Terms of Service | Security Statement | Employment