Activation Patterns Associated with Explicit and Implicit Processing

C ATHERINE H ANSON AND S TEPHEN J OS E´ H ANSON RUMBA Labs, Rutgers University, Newark, NJ 07102, USA

Background

Behavioral Results

Interactivity Between Areas

A dissociation between what appears to be intentional, explicit processing of information and that which relies on incidental, implicit processing has been well documented for some time (Schacter, 1987). For example, amnesics who are unable to consciously recall presented material, will be able to demonstrate memory for the material when tested indirectly (e.g., with priming or stem completion tasks). In healthy individuals, manipulations that affect performance on one type of memory task may have little effect on performance of the other type of task. For example, whereas changes in the surface features of presented items (e.g., font, color) will affect implicit memory, this manipulation often has little effect on explicit memory.

Accuracy: In general, participants who had encoded material implicitly made fewer errors on both retrieval tasks than did participants who had encoded material explicitly, F(1,44)=11.81, p<.01. Participants in the implicit retrieval condition also were significantly more accurate than those in the explicit retrieval condition, F(1,44)=5.46, p<.05.

Encoding Common areas that were recruited across both conditions of the encoding phase include: cingulate, superior temporal gyrus, and lentiform. These areas are associated with cognitive control, visual search (superior temporal gyrus), and learning (lentiform). In addition to these areas, the explicit encoding condition appears to recruit inferior parietal lobule, an area associated with visuotemporal attention. The implicit encoding condition uniquely recruited inferior frontal gyrus, an area associated with attentional selection (Zhang et al., 2004).

One explanation for the dissociations observed between implicit and explicit memory is that the performance differences reflect distinct memory systems (Schacter& Tulving, 1994). An alternative explanation is that dissociations between implicit and explicit tasks arise, not because of distinct memory systems, but because different tasks require different types of processing. Whereas explicit tasks tend to tap into conceptual processing, implicit tasks may be more data-driven (Roediger, 1990). Some of the difficulty in evaluating the two models can be attributed to the ambiguity surrounding the terms explicit and implicit. It is not always clear whether these terms are being used to refer to types of tasks, cognitive states, or theoretical constructs (e.g., memory systems).

Study Goals

Response Time: No main effects were found for either encoding or retrieval, however, a significant interaction was obtained, F(1,44)=7.24, p<.01.

main effect (explicit encoding) significant, p<.01

interaction significant, p<.02

GLM Analysis Analysis was carried out using FEAT (FMRI Expert Analysis Tool) Version 5.4, part of FSL (FMRIB’s Software Library, www.fmrib.ox.ac.uk/fsl). Shown here are the zstat maps averaged across participants in a given condition.

• examine explicit and implicit retrieval as a function of initial encoding • use fMRI to look at the neural correlates of explicit and implicit processing • produce process models of explicit and implicit processing

Method Participants: Forty-eight Rutgers undergraduates participated for course credit or cash. Participants were randomly assigned in equal number to the four experimental conditions. explicit encoding explicit retrieval

explicit encoding implicit retrieval

implicit encoding explicit retrieval

implicit encoding implicit retrieval

Stimuli & Procedure:

Explicit encoding activation

Implicit encoding activation

Explicit ret, explicit enc

Implicit ret, explicit enc

Explicit ret,implicit enc

Implicit ret, implicit enc

Retrieval Cingulate appears again during retrieval and is, in fact, the only area that appears in all four retrieval conditions. Cingulate was found to have both input and output connections to other areas in all but the explicit encoding, explicit retrieval condition. Inferior parietal lobule appears in all conditions with the exception of explicit retrieval following implicit encoding. In retrieval conditions following explicit encoding inferior parietal lobule was interactive, both receiving and producing input. This contrasts with the role of inferior parietal lobule during implicit retrieval following implicit encoding where inferior parietal lobule exerted a strong influence on cingulate, but did not receive significant input from other areas. Recent work (Reber et al, 2005) suggests that fusiform gyrus, an area often associated with face recognition, may play a role in priming. It was found in explicit retrieval following implicit encoding. Lentiform, an area associated with learning and possibly executive cognitive function (Brown et al., 1957) was found to interact with cingulate in retrieval conditions following implicit encoding, but not during implicit retrieval. Note also that supramarginal gyrus, an area associated with language processing, appears only in the explicit retrieval condition following explicit encoding.

ENCODING CONDITION explicit encoding: Will you remember this word? implicit encoding: Is the sentence true? RETRIEVAL CONDITION explicit retrieval: Did you see this word before? implicit encoding: Is the letter string a word? Encoding Phase During the encoding phase participants in the explicit condition saw either a series of words (e.g., dog, phone, cloud) and were asked to decide for each word if they thought they would be able to remember it. In the implicit encoding condition participants saw sentences containing the words presented to participants in the explicit encoding condition and were asked to judge the truth value of each sentence. Retrieval Phase Participants in the explicit retrieval condition were presented with a series of words that either had been seen previously seen and not and were asked to judge which had been seen during the encoding phase. Participants in the implicit retrieval condition were presented with a series of letter strings (words and nonwords) and asked to judge if a given string was a word.

General Conclusions

RUMBA Causal Modelling Method FSL analysis typically produced 10-12 areas at p<.05 (cluster level). Subsequent clustering (see POSTER #617) was performed in which both subject agreement (number of subjects that had voxels in similar areas >50%) and a significant number of voxels (>50) were used as criteria. This clustering yielded six areas: cingulate, superior temporal gyrus, lentiform, inferior parietal lobule, inferior frontal gyrus, and supermarginal gyrus. Starting with Clusters

Extract Timeseries

Filter and Average

Compute Covariance

• differences in performance are consistent with transfer appropriate processing (Blaxton, 1989) • brain areas are recruited differentially as a function of processing demands • unlikely that distinct memory systems underlie explicit or implicit processing

References • Blaxton, T. A. (1989). Investigating dissociations among memory measures: Support for a transfer-appropriate processing framework. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15, 657-668. • Brown, L.L., Schneider, J.S., Lidsky, T.I. (1997). Current Opinion in Neurobiology, 7, 157-163. • Reber, P.J., Gitelman, D.R., Parrish, T.B., and Mesulami, M. (2005). Cerebral Cortex, 15(6):787-795. • Roediger, H.L. (1990). Implicit memory. American Psychologist, 45, 1043-1056 • Schacter, D.L. (1987). Implicit memory: History and current status. Journal of Experimental Psychology: Learning, Memory, and Cognition, 13, 501-518. • Schacter, D.L., and Tulving, E. (1994). What are the memory systems of 1994? In D.L. Schacter and E. Tulvings (Eds.), Memory Systems, Cambridge, MA: MIT Press

Covariance Matrix

Structural Equation Modeling

• Zhang, J.X., Feng, C., Fox, P.T., Gao, J., and Tan, L.H. (2004). Is left inferior frontal gyrus a general mechanism for selection? Neuroimage, 23, 596.603.

Acknowledgements Searched all possible graphs (5 nodes: 54,021 graphs)

Supported by NSF (EIA-0205178), and the James S. McDonnell Foundation Thanks to Yarolsav Halchenko, Donovan Elflord, and Toshihiko Matsuka