Limits...
Computational models of the Posner simple and choice reaction time tasks.

Feher da Silva C, Baldo MV - Front Comput Neurosci (2015)

Bottom Line: In doing so, main findings of experimental research on RT were replicated: the relative frequency effect, suboptimal RTs and significant error rates due to noise and invalid cues, slower RT for choice RT tasks than for simple RT tasks, fastest RTs for valid cues and slowest RTs for invalid cues.Analysis of the optimized systems revealed that the employed mechanisms were consistent with related findings in neurophysiology.Our models, however, reveal that the main problems that must be overcome to perform the Posner task effectively are distinguishing signal from external noise and selecting the appropriate response in the presence of internal noise.

View Article: PubMed Central - PubMed

Affiliation: Department of General Physics, Institute of Physics, University of São Paulo São Paulo, Brazil.

ABSTRACT
The landmark experiments by Posner in the late 1970s have shown that reaction time (RT) is faster when the stimulus appears in an expected location, as indicated by a cue; since then, the so-called Posner task has been considered a "gold standard" test of spatial attention. It is thus fundamental to understand the neural mechanisms involved in performing it. To this end, we have developed a Bayesian detection system and small integrate-and-fire neural networks, which modeled sensory and motor circuits, respectively, and optimized them to perform the Posner task under different cue type proportions and noise levels. In doing so, main findings of experimental research on RT were replicated: the relative frequency effect, suboptimal RTs and significant error rates due to noise and invalid cues, slower RT for choice RT tasks than for simple RT tasks, fastest RTs for valid cues and slowest RTs for invalid cues. Analysis of the optimized systems revealed that the employed mechanisms were consistent with related findings in neurophysiology. Our models predict that (1) the results of a Posner task may be affected by the relative frequency of valid and neutral trials, (2) in simple RT tasks, input from multiple locations are added together to compose a stronger signal, and (3) the cue affects motor circuits more strongly in choice RT tasks than in simple RT tasks. In discussing the computational demands of the Posner task, attention has often been described as a filter that protects the nervous system, whose capacity is limited, from information overload. Our models, however, reveal that the main problems that must be overcome to perform the Posner task effectively are distinguishing signal from external noise and selecting the appropriate response in the presence of internal noise.

No MeSH data available.


Related in: MedlinePlus

Median RTs in Experiment 2. Median RTs for SRT and CRT tasks. In Conditions A and B (top two rows), results were from generation 30 for SRT tasks and from generation 50 for CRT tasks. In Condition C (bottom row), all results are from generation 300. N = 50. RTs are given in the model's arbitrary time units.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4488626&req=5

Figure 7: Median RTs in Experiment 2. Median RTs for SRT and CRT tasks. In Conditions A and B (top two rows), results were from generation 30 for SRT tasks and from generation 50 for CRT tasks. In Condition C (bottom row), all results are from generation 300. N = 50. RTs are given in the model's arbitrary time units.

Mentions: For Condition A, the results, displayed in the top row of Figure 6, show that RTs decreased as the populations of neural networks evolved. By generation 300, RT was 1 time unit for all cue types. This is the minimum RT in correct trials, since the simulation advances in discrete time steps. At generation 0, the error rate (the sum of anticipated, slow, and incorrect response rates) was 100%, but from generation 30 onward, for the SRT experiment, and from generation 80 onward, for the CRT task, error rates were 0%. Thus, the neural networks achieved optimal performance, responding to the target at the earliest possible instant without errors. The top row of Figure 7 is a snapshot of the results before optimal performance was achieved, taken from generation 30 for SRT tasks and from generation 50 for CRT tasks. RTs were around 2–3 time units and error rates were below 2.5%. At that early point, RTs were shortest for valid cues and longest for invalid cues (the difference is statistically significant for all comparisons except those including the neutral cue in the SRT task, at the level of α = 0.05, with N = 50). The difference between RTs for valid and neutral cues might have been, however, due to stimulus frequency, as discussed above.


Computational models of the Posner simple and choice reaction time tasks.

Feher da Silva C, Baldo MV - Front Comput Neurosci (2015)

Median RTs in Experiment 2. Median RTs for SRT and CRT tasks. In Conditions A and B (top two rows), results were from generation 30 for SRT tasks and from generation 50 for CRT tasks. In Condition C (bottom row), all results are from generation 300. N = 50. RTs are given in the model's arbitrary time units.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4488626&req=5

Figure 7: Median RTs in Experiment 2. Median RTs for SRT and CRT tasks. In Conditions A and B (top two rows), results were from generation 30 for SRT tasks and from generation 50 for CRT tasks. In Condition C (bottom row), all results are from generation 300. N = 50. RTs are given in the model's arbitrary time units.
Mentions: For Condition A, the results, displayed in the top row of Figure 6, show that RTs decreased as the populations of neural networks evolved. By generation 300, RT was 1 time unit for all cue types. This is the minimum RT in correct trials, since the simulation advances in discrete time steps. At generation 0, the error rate (the sum of anticipated, slow, and incorrect response rates) was 100%, but from generation 30 onward, for the SRT experiment, and from generation 80 onward, for the CRT task, error rates were 0%. Thus, the neural networks achieved optimal performance, responding to the target at the earliest possible instant without errors. The top row of Figure 7 is a snapshot of the results before optimal performance was achieved, taken from generation 30 for SRT tasks and from generation 50 for CRT tasks. RTs were around 2–3 time units and error rates were below 2.5%. At that early point, RTs were shortest for valid cues and longest for invalid cues (the difference is statistically significant for all comparisons except those including the neutral cue in the SRT task, at the level of α = 0.05, with N = 50). The difference between RTs for valid and neutral cues might have been, however, due to stimulus frequency, as discussed above.

Bottom Line: In doing so, main findings of experimental research on RT were replicated: the relative frequency effect, suboptimal RTs and significant error rates due to noise and invalid cues, slower RT for choice RT tasks than for simple RT tasks, fastest RTs for valid cues and slowest RTs for invalid cues.Analysis of the optimized systems revealed that the employed mechanisms were consistent with related findings in neurophysiology.Our models, however, reveal that the main problems that must be overcome to perform the Posner task effectively are distinguishing signal from external noise and selecting the appropriate response in the presence of internal noise.

View Article: PubMed Central - PubMed

Affiliation: Department of General Physics, Institute of Physics, University of São Paulo São Paulo, Brazil.

ABSTRACT
The landmark experiments by Posner in the late 1970s have shown that reaction time (RT) is faster when the stimulus appears in an expected location, as indicated by a cue; since then, the so-called Posner task has been considered a "gold standard" test of spatial attention. It is thus fundamental to understand the neural mechanisms involved in performing it. To this end, we have developed a Bayesian detection system and small integrate-and-fire neural networks, which modeled sensory and motor circuits, respectively, and optimized them to perform the Posner task under different cue type proportions and noise levels. In doing so, main findings of experimental research on RT were replicated: the relative frequency effect, suboptimal RTs and significant error rates due to noise and invalid cues, slower RT for choice RT tasks than for simple RT tasks, fastest RTs for valid cues and slowest RTs for invalid cues. Analysis of the optimized systems revealed that the employed mechanisms were consistent with related findings in neurophysiology. Our models predict that (1) the results of a Posner task may be affected by the relative frequency of valid and neutral trials, (2) in simple RT tasks, input from multiple locations are added together to compose a stronger signal, and (3) the cue affects motor circuits more strongly in choice RT tasks than in simple RT tasks. In discussing the computational demands of the Posner task, attention has often been described as a filter that protects the nervous system, whose capacity is limited, from information overload. Our models, however, reveal that the main problems that must be overcome to perform the Posner task effectively are distinguishing signal from external noise and selecting the appropriate response in the presence of internal noise.

No MeSH data available.


Related in: MedlinePlus