Cohen JD, Dunbar K, McClelland JL. On the control of automatic processes: A parallel distributed processing account of the Stroop effect. Psychological Review 1990; 97: 33261.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....j (t) of the previous layer and a natural exponential decay of activation depending on some decay constant l. This can be rewritten as: Sum i (t) l w ij l Prev j (t) 1 l )Sum i (t 1) j which with l t and w l w corresponds exactly with the equivalent formulation of Cohen et al. [5]. In the asymptotic state Sum i (t) Sum i (t 1) so: Sum (t) w ij l Prev (t) j . It follows that the asymptotic state of our cascaded network is the same as a standard feedforward network with weights w l. Assuming the right way to train the cascading network is to adjust the weights ....

Cohen, J., Dunbar, K. & McClelland (1990). On the control of automatic processes: A parallel distributed processing model of the Stroop task, Psychological Review, 97, 332-361.

....Monsell et al. 1992; Rastle and Coltheart, 1999) What has been lacking from these studies is a computational account of strategic control that is integrated with a broader theory of reading and cognition. The current study investigates a general mechanism of strategic control, termed input gain (Cohen et al. 1990; Kello and Plaut, 2000; Kello et al. 2000) Our investiRequests for reprints should be addressed to Christopher T. Kello (kello cats.ucsc.edu) Psychology Department, University of California, Santa Cruz CA, 95064. The research was supported by the House Ear Institute and an NIMH FIRST award ....

....processing in the entire system, rather than a subset of its components. However, if input gain is controlled over individual system components, it can serve to enhance or diminish the contributions of those components to processing (for a demonstration in the context of Stroop interference, see Cohen et al. 1990). Certain word reading tasks may be best performed by enhancing or diminishing the contributions from either semantic or phonological knowledge. We leave it to future research to examine whether the findings argued to be in favor of route emphasis could be accommodated by controlling input gain ....

Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On the control of automatic processes: A parallel distributed processing account of the Stroop effect. Psychological Review, 97(3), 332-361.

....properties of the data. Traditionally, psychological models have aimed for depth over breadth, and the cost has been that a model of one phenomenon, say the word superiority effect (McClelland Rumelhart, 1981) may have little in common with a model of some other phenomenon, say the Stroop task (Cohen, Dunbar, McClelland, 1990), even though the two models are ostensibly of the same fundamental process, reading in this case. 18 In model building, the distinction between old and new data is seldom clear. One often constructs the model with particular data in mind, and then discovers that the model, with no or minor ....

Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On the control of automatic processes: A parallel distributed processing model of the Stroop effect. Psychological Review, 97, 332--361.

....in another. Skill, learning and dyslexia June 12, 1995 16 P6. Between module connections are bidirectional and excitatory, so that processing is interactive. P7. Within module connections are bidirectional and inhibitory, so that processing is competitive. In an earlier paper, Cohen et al. 1990) demonstrated that a simple feedforward model with one hidden layer, using the back propagation learning algorithm (Rumelhart et al. 1986) could simulate many of the key empirical findings of timing in the Stroop task (Stroop, 1935) In subsequent work Cohen et al. 1992) adapted the network to ....

....can be applied to the Pacman task. Here, though, there is a complication in that the subject must both learn the finger movement mappings and also the optimal time to initiate each turn. The learning of the finger movement mappings involves a consciously mediated learning phase, referred to by Cohen et al. 1990) as indirect learning and by Anderson (1983) as declarative learning, which is beyond the scope of a PDP analysis. Following Cohen et al. we shall ascribe it to an indirect module needed in the initial stages of learning but gradually being replaced by a direct mapping to the appropriate ....

[Article contains additional citation context not shown here]

Cohen, J.D., Dunbar, K. & McClelland, J.L. (1990). On the control of automatic processes: A parallel distributed processing model of the Stroop effect. Psychological Review, 97, 332361.

....countless experiments were designed to assess whether it is possible to establish that a given process is entirely automatic. Crucially for my argument, automaticity was described as a dichotomous and binary property. A given process, thus, was thus thought to be either automatic or controlled. Cohen, Dunbar McClelland (1990), along with others before them (e.g. Logan, 1980; Kahneman Chajczyk, 1983; McLeod and Dunbar, 1988) attacked this position and claimed that automaticity is really a continuous dimension. Illustrative empirical arguments can be found in a study by McLeod and Dunbar (1988) in which participants ....

....early in training, but produced an increasingly large slowdown with practice at the shape naming task. Clearly, then, one can obtain continuous interference effects, the magnitude of which depends on the amount of training that each dimension has been allowed to benefit from. The modeling work of Cohen et al. 1990) showed that the effects of practice can be simply expressed as the strength of a processing pathway in a connectionist model. The model was successful in accounting for McLeod and Dunbar s data, and suggests that automaticity, rather than being dichotomous, is best thought of as a continuous ....

Cohen, J.D., Dunbar, K., & McClelland, J.L. (1990). On the control of automatic processes: A parallel distributed account of the Stroop effect. Psychological Review, 97, 332--361.

....apparatus which, in humans, imposes a floor effect as response latencies and durations approach their maximum speeds. This issue is peripheral to our research question, so we did not address it here (for an additional explanation of the difference between Stroop facilitation and interference, see Cohen, Dunbar, McClelland, 1990). Error analyses with low gain. Figure 10 graphs the mean error rates, and effects of interference and facilitation, as a function of congruency and SOA. The pattern of error rates followed the latency results in the model, and this matches the pattern of results found in Experiment 1. In ....

....stimuli are presumably represented separately at one or more levels of stimulus processing in the human cognitive system. Instructions to name the color and not the word (or vice versa) would privilege the processing of color stimuli over word stimuli via some type of attentional mechanism (see Cohen et al. 1990). When both stimuli are presented simultaneously (as in the 0 ms SOA condition) this attentional mechanism is able to effectively quash the influence of the interfering word. However, when the interfering word is presented after the target color, attention is drawn to the onset of a new visual ....

Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On the control of automatic processes: A parallel distributed processing account of the Stroop effect. Psychological Review, 97, 332--361.

....of patient behaviour. The computational models provide a complementary methodology to investigate assumptions made explicit by theories of normal and impaired cognitive processing. In this project we present a study of Cohen, Dunbar and McClelland s (CDM) neural network model of the Stroop effect [1] subjected to different damage techniques. We compare the resulting behaviour with results from the literature on the performance of patients with Dementia of the Alzheimer s Type (DAT) 1.1 Modelling Automaticity and Control in Alzheimer s Disease Many types of skilled activities such as ....

....damage conditions were compared to those of the DAT individuals and the older control subjects. 2.1.1 Replication: discovering the effects of t 0 The first stage of our study involved simulating the original CDM network. The simulation parameters and procedures were as reported in Cohen et al. [1], except for two aspects: a) The standard deviation of the noise in the evidence accumulators was set to 0.01 (instead of 0.1, correcting an error noted in [9, p. 874 footnote 1] and (b) a second timing parameter, t 0 was introduced: Cohen et al. 1] reported allowing the network to settle ....

[Article contains additional citation context not shown here]

Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On the control of automatic processes: A parallel distributed processing account of the Stroop effect. Psychological Review, 97, 332-361.

....actually the adjustments to the connections in a distributed connectionist system rather than stored instances. McClelland and Rumelhart (1985) for example, showed how a simple superpositional system can capture several patterns of data previously taken as supporting instance based theories, and Cohen, Dunbar, and McClelland (1990) demonstrated that distributed connectionist models trained with back propagationcan capture the power law of practice just as Logan s instance models do. It seems somewhat more plausible to us that multiple occurrences of a meaningful cognitive entity such as a letter or word might be mapped ....

Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On the control of automatic processes: A parallel distributed processing account of the Stroop effect. Psychological Review, 97(3), 332-361.

No context found.

Cohen JD, Dunbar K, McClelland JL. On the control of automatic processes: A parallel distributed processing account of the Stroop effect. Psychological Review 1990; 97: 33261.

No context found.

Cohen, J.D., Dunbar, K., and McClelland, J.L. (1990) On the control of automatic processes: a parallel distributed processing account of the Stroop effect. Psychol. Rev. 97, 332--361.

No context found.

Cohen, J.D. et al. (1990) On the control of automatic processes: a parallel distributed processing account of the Stroop effect. Psychol. Rev. 97, 332--361

No context found.

J.D. Cohen, K. Dunbar, and J.L. McClelland. On the control of automatic processes: A parallel distributed processing account of the stroop effect. Psychological Review, 97(3):332--361, 1990.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC