Information Technology Reference
In-Depth Information
Table 2.
Parameters for predicting BOLD response of five modules
Module Visual Goal Retrieval Imaginal Production
m
0.8
1
1
1.3
0.1
Parameters
s
1.5
0.9
2.5
1.5
1.5
b
6
7
2
6.7
6
related to heuristics, participants make a number of judgments by the state of
goal and retrieve heuristics for the answer and finally make their responses. We
simplify the processes in one sentence: only encoding valid information in related
dimensions for heuristics retrieval by goal status oriented. In the model, ways of
problem presentation, complexity of heuristics and status of goal take important
roles in the heuristics retrieval.
3R su s
3.1 Behavioral Results
Behavioral results, the mean latencies (also called response time (RT)) over 16
participants from the presentation of tasks stimulus to the thumb press are 1.53
seconds in simple tasks and 2.90 seconds in complex tasks. These results came
from behavioral data collected synchronous with fMRI scanning. The means RT
are for correct trials only. The difference between simple and complex tasks is
1.37 seconds which shows that the heuristics application in complex tasks cost
more time because visual encoding, memory retrieval, problem representation,
goal control and production rules are more complex than simple ones. All of these
cognitive elements are involved in the retrieval and application of heuristics.
3.2 fMRI BOLD Effect Results
Dotted lines in Figure 3 show the BOLD response of five predefined brain regions
for simple and complex 4*4 Sudoku tasks. These results show that complex
tasks always have larger time scale and steeper shape than the simple ones. The
complexity of heuristics makes these differences. More explanation will be given
later.
3.3
Fitting an ACT-R to the Behavioral Results
The predictions of behavioral results RT are 1.50 seconds in simple tasks and 2.84
seconds in complex tasks. The correlation of latency of ACT-R theory and real
RT data is 0.999, and the difference is 0.04. The well fitness of ACT-R theory to
participants data in behavioral results demonstrates that the modeling cognitive
processes mentioned in Section 2, are a certain extent reasonable. Tables 3 and
4 show the information processing processes on five modules of ACT-R for the
two kinds 4*4 Sudoku based on trace outputs of ACT-R model.
Visual Module
. For both tasks, participants need to find out '?' firstly for
retrieval and then retrieve digits on related dimensions. For the simple ones, the
