Nice guys finish fast and bad guys finish last: Facilitatory vs. inhibitory interaction in parallel systems

Abstract

Systems Factorial Technology is a powerful framework for investigating the fundamental properties of human information processing such as architecture (i.e., serial or parallel processing) and capacity (how processing efficiency is affected by increased workload). The Survivor Interaction Contrast (SIC) and the Capacity Coefficient are effective measures in determining these underlying properties, based on response-time data. Each of the different architectures, under the assumption of independent processing, predicts a specific form of the SIC along with some range of capacity. In this study, we explored SIC predictions of discrete-state (Markov process) and continuous-state (Linear Dynamic) models that allow for certain types of cross-channel interaction. The interaction can be facilitatory or inhibitory: one channel can either facilitate, or slow down processing in its counterpart. Despite the relative generality of these models, the combination of the architecture-oriented plus the capacity oriented analyses provide for precise identification of the underlying system.

Parallel interactive retrieval of item and associative information from event memory

“…For example, while the analytic form of the SIC is known for some classes of coactive system (Townsend & Nozawa, 1995;Fific et al, 2008;, there as yet is no general proof that covers all coactive systems, let alone those with more complex interactions of the kind we considered. While the converging evidence from our various analyses and their relation to other results in the literature (as discussed both above and below) puts our conclusions regarding memory on firm footing, our efforts also contribute to the ongoing project to understand the qualitative signatures of different cognitive systems and to develop experimental means to detect them (Townsend & Wenger, 2004;Eidels et al, 2011;Little et al, 2011;Yang et al, 2014).…”

“…In addition, error rates tend to be higher in our study than in those to which SFT is typically applied. We therefore obtained predictions for the SFT measures using a combination of simulations (Diederich & Busemeyer, 2003;Fific et al, 2008;Eidels et al, 2011) and analytic methods (Townsend & Thomas, 1994), as described in Appendix A. The resulting predictions allow for both a large proportion of errors (from near 0% to 96%) and violations of selective influence; while we lose the statistical power available with low error rates and selective influence, we can be confident that we have provided a fair qualitative characterization of each possible retrieval architecture.…”

“…For example, if strongly encoded item information makes it easier to encode associative information (or vice versa), this would constitute a facilitatory interaction at storage that would be separate from any facilitation that might occur as a function of the architecture of the retrieval processes. Eidels et al (2011) termed this a distinction between "pre-accumulator" and "post-accumulator" interactions, where "pre-accumulator" in our case applies to processes like storage that occur prior to retrieval while "postaccumulator" applies to the retrieval processes themselves. These two kinds of interaction are explicitly disentangled within our LBA framework, which separates retrieval architecture-"post-accumulator" interactions that result from pooling information across accumulators-from the inputs to that architecture-the drift rates that reflect the evidence retrieved from memory as well as any "pre-accumulator" interactions.…”

“…Because we could not guarantee that our repetition manipulations would selectively influence item and associative strength, we augmented analytical predictions for the SIC (Townsend & Nozawa, 1995) with simulations that examine the consequences of pooling between the two processes (for additional simulation results in high-accuracy situations, see Eidels et al, 2011).…”

Numerical predictions for serial, parallel, and coactive logical rule-based models of categorization response time

“…This approach has achieved considerable success in answering fundamental questions about about whether multiple stimulus dimensions are processed sequentially in a serial fashion or simultaneously in parallel, or are pooled into a single, coactive processing channel Miller, 1982;Townsend & Wenger, 2004a, 2004b; see also Eidels, Houpt, Altieri, Pei, & Townsend, 2011;Palmer & McLean, 1995;Thornton & Gilden, 2007).…”