Reaction times

Question 1

Cognitive neuroscience studies what?

Answer 1

Internal representations and the operations/transformations on them. What happens in the black box between input and output of information.

Question 2

Example RT study: Sternberg recognition memory task. 4 letters, 1 letter presented. Respond whether stimulus was in target string yes/no. (1) RQ?

Answer 2

Is recognition memory a parallel process whereby the cue is compared with all target letters at the same time or is it serial, comparing them one by one?

Question 3

Example RT study: Sternberg memory search / recognition memory task. 1 to 8 letters, 1 letter presented. Respond whether stimulus was in target string yes/no. (2) Method, prediction, results?

Answer 3

If recognition memory a serial process, RT would show a linear increase with the number of items, whereas if it was parallel it would not show this. -> this is what happened.
RT = constant + 40(ms)*number of items

Question 4

Example parallel processes? Hint: word superiority effect / Reicher 1969

Answer 4

When flashing either XAXX / RACK / KARC very briefly, and the task is to respond as accurately as possible which of two targets A or E is in the target string, the word RACK has best performance the word level is processed parallel to the singular letters, and because both offer info as to whether A is in target, accuracy is higher.

Question 5

Example RT study: Posner 1986 investigated RT in a letter matching task. He presented a probe letter and a target letter, pps had to decide whether probe is in same category (vowels/consonants) as target. From RT data they concluded that we have at least 3 different representations of letters. How/why?

Answer 5

All 5 conditions had different RTs: A - A (physical identity)
A - a (phonetic identity)
A - E (same categorical identity)
C - S (same categorical identity)
A - S (different categorical identity)

Question 6

In the Sternberg recognition memory study, they also gained insight on the order of the processing stages “encoding”, “comparison”, “decision”, “responding”. Hint: exhaustive vs self terminating search

Answer 6

If self terminating search, expectation ‘yes, present in target’ would show RT curve 50% lower than in case of ‘no’ answers, whereas exhaustive search would show the same. The pattern of exhaustive search was found, meaning that the letter is compared with every letter in target serially, and only at end a decision is made instead of comparing and deciding at every letter.

Question 7

RT definition: operational and theoretical

Answer 7

Operational: time between stimulus onset and an overt response
Assumption: pp wants to respond as accurately and as fast as possible.
Theor: min. time needed for correct response.

Question 8

Advantages RT?

Answer 8

• susceptible to subtle condition differences, orderly data that can easily be used to make models
• ratio lvl: meaningful 0 point and linear differences (i.e. 400 - 300 means the same as 300 - 200)
• not arbitrary like other psy constructs
• applicable to real life (reflects actual performance)

Question 9

How can different RTs be interpreted as resembling different processes?

Answer 9

Processes inherent in different pathways in brain, which means longer or shorter travelling time for signals, resulting in different RTs

Question 10

Limitations RT

Answer 10

• mental processes not observed directly
• can only study simple processes bc too complex processes leaves too much room for individ differences in strategy used. (rule: 200 - 1000 ms)

Question 11

Donders subtraction main idea

Answer 11

To find out duration of a task, subtract RT from task B which only contains 1 extra process but is the same otherwise, from task A

Question 12

Subtraction method 3 types and processes involved in each

Answer 12

A Simple RT: 1 stimulus, 1 response (sensory, motor)
B 2-choice: 2 poss. stim in 1 trial, each demanding diff response (sensory, discriminatory, response selection, motor)
C Go no go / disjunctive RT: 2 stim, 1 response (sensory, discriminatory, motor)

Question 13

One problem with subtraction method is that it relies on strong assumptions. Which three are there?

Answer 13

1. seriality: stages are succesive, not parallel, so that durations sum up
2. pure insertion: adding or deleting 1 stage does not influence duration of another (no interaction)
3. accurate preknowledge of structure processes in task: we need to know how many and which stages involved in order to make good comparison tasks

Question 14

Critique of pure insertion assumption

Answer 14

• Often insertion not pure, e.g. inserting a response selection stage affects duration of motor process (as measured by ERPs - overt response)
• Go no go / disjunctive Donders task also requires a decision / response selection stage so cant be purely deleted from 2 choice task to calculate duration motor execution.

Question 15

Additive Factor Method what is it

Answer 15

1. Factorial design (at least 2 factors) (can also compare 2 groups but does not have to)
2. Measure RT in all different combinations
3. if (sig) interaction between 2 factors, they arent additive and thus are in same stage, otherwise additive and in different stage

Question 16

AFM main propositions, how is it better, which assumptions are solved and which remain.

Answer 16

• also SUCCESIVE stages: each stage performs transformation on input, then outputs to next.
• manipulations that change IV impact its duration but not its output
• solves preknowledge problem bc it actually tells you how many independent stages
• pure insertion is not necessary

Question 17

How was an altered Sternberg memory search task used with AFM? Factorial design: stim degradation (yes/no) vs number of items. RQ and prediction

Answer 17

• prediction: if identification before comparison, then degraded stimulus condition should only have a higher intercept but same slope for number of items
• whereas if same stage (immediately compare degraded stim to each target letter), same intercept but steeper slope as it takes longer to compare degraded stim with each target letter.

Question 18

How was an altered Sternberg memory search task used with AFM? Results

Answer 18

Additive, higher intercept but same slope, first ID then compare.

Question 19

Limitations AFM

Answer 19

• Since H0 = no interaction, being able to conclude that stages are additive depend upon power and sample size.
• Requires stage robustness: additive shouldnt change once a third factor is added (eg same test but now in different age group)
• Interaction doesnt always mean same stage: can also occur if diff stimulus features are processed in parallel (smulders p 14)
• cant conclude about order or content of stages

Question 20

What are ways to deal with speed accuracy tradeoffs?

Answer 20

• give clear instructions to pps whether to focus on fastest response, most accurate response, or both
• remove trials in which an error was made (RT not valid)
• comparing conditions on error percentage (position on tradeoff curve)
• measure along the entire curve
• mathematically model RT and accuracy in diffusion model to estimate position on tradeoff curve

Question 21

What is the diffusion model

Answer 21

• Assumes information accumulates continuously from onset to response, in direction of decision1 or decision2, with random noise that turns counter up or down -> explains trial to trial variability RT.
• Normal distribution around WHEN decision is made but with a constant CRITERION of how much info needed to make decision (counter).
• By estimating paramaters, the RT can be modelled until it matches the pattern observed, and you can use the parameters (including RT/criterion).

Question 22

How to deal with outliers?

Answer 22

• may remove outliers in lvl 1 (within subject) but removing an outlier in lvl 2 (compared with condiiton) is discouraged and must be well motivated
• practice rounds with feedback -> reduce intercept but not slope RT

Question 23

How was RT used to investigate info processing in depression
2 stages: perceptual processing / response selection
Azorin

Answer 23

depression: stimulus intensity of light cue was varied & response compatibility (press right if cue right vs press left). vs control group, perceptual processing was unaffected (main effect of stim intensity, no stim intensity * condition interaction)

Question 24

How was RT used to investigate info processing in aging?
3 stages: perceptual processing / response selection / motor adjustment
Bonin Guillaume

Answer 24

besides stim intensity and response compatibility, motor adjustment stage was investigated with an early vs late cue. early cue leads to longer RT.
this effect was larger for old ppl than young (interaction of cue timing * age)

Question 25

Bonin Guillaime 2004 are interested if depression is the same for old and young ppl. Bc of confounding reasons, they could not compare old vs young patients, but instead compared old nondepressed ppl vs old depressed ppl. Which result hinted that the structure of processing stages is the same in old depressives as in young depressives?

Answer 25

Main effects of depression, and all 3 task variables. There was no interaction between task variables (meaning in these groups, they are additive and 3 diff stages, which means structure is the same as normal)

Question 26

Bonin Guillaime 2004 are interested if depression is the same for old and young ppl. Bc of confounding reasons, they could not compare old vs young patients, but instead compared old nondepressed ppl vs old depressed ppl. Which 2 significant interactions showed the impairing effect of depression in this group? Hint: 1 was in RT, the other was in errors.

Answer 26

RT: sig interaction depressionmotor adjustment: depressed groups was slowed more than cntrls by early cue.
Errors: sig interaction depressionresponse selection: incompatible responses lead to more errors for depressed than nondepressed, meaning they may have done a speed accuracy tradeoff.

Question 27

What are some modern improvements of the subtraction paradigm?

Answer 27

• probability of giving a response is kept approx equal across trials
• response preparation is kept approx equal across conditions