Remember-Know: A Matter of Confidence

Combined pharmacological and electrophysiological dissociation of familiarity and recollection

Dual-process theories of recognition memory hypothesize separate underlying familiarity and recollection processes, but the necessity of multiple processes is debated. Previous research has suggested that scalp-recorded event-related brain potentials (ERPs) may index the activity of separate familiarity and recollection processes. Other research indicates that the amnestic drug midazolam impairs recollection more than familiarity. Here, we used a convergent pharmacological and electrophysiological approach to manipulate and monitor human brain activity and provide evidence for separate processes. Midazolam selectively influenced the putative ERP-correlate of recollection but not the putative ERP-correlate of familiarity. Under control conditions (saline), subjects' accuracy correlated with the recollection-related but not the familiarity-related ERP component, suggesting that recollection was dominant in driving memory. The opposite pattern was observed under midazolam administration, suggesting that when recollection fails, subjects may leverage familiarity to compensate. Thus, in contrast to perspectives holding that familiarity represents the default process, these results suggest that recollection was dominant until its impairment unveiled the influence of familiarity.

Cognitive neuroimaging: Cognitive science out of the armchair

Cognitive scientists were not quick to embrace the functional neuroimaging technologies that emerged during the late 20th century. In this new century, cognitive scientists continue to question, not unreasonably, the relevance of functional neuroimaging investigations that fail to address questions of interest to cognitive science. However, some ultra-cognitive scientists assert that these experiments can never be of relevance to the study of cognition. Their reasoning reflects an adherence to a functionalist philosophy that arbitrarily and purposefully distinguishes mental information-processing systems from brain or brain-like operations. This article addresses whether data from properly conducted functional neuroimaging studies can inform and subsequently constrain the assumptions of theoretical cognitive models. The article commences with a focus upon the functionalist philosophy espoused by the ultra-cognitive scientists, contrasting it with the materialist philosophy that motivates both cognitive neuroimaging investigations and connectionist modelling of cognitive systems. Connectionism and cognitive neuroimaging share many features, including an emphasis on unified cognitive and neural models of systems that combine localist and distributed representations. The utility of designing cognitive neuroimaging studies to test (primarily) connectionist models of cognitive phenomena is illustrated using data from functional magnetic resonance imaging (fMRI) investigations of language production and episodic memory.

Models of recognition: a review of arguments in favor of a dual-process account

The majority of computationally specified models of recognition memory have been based on a single-process interpretation, claiming that familiarity is the only influence on recognition. There is increasing evidence that recognition is, in fact, based on two processes: recollection and familiarity. This article reviews the current state of the evidence for dual-process models, including the usefulness of the remember/know paradigm, and interprets the relevant results in terms of the source of activation confusion (SAC) model of memory. We argue that the evidence from each of the areas we discuss, when combined, presents a strong case that inclusion of a recollection process is necessary. Given this conclusion, we also argue that the dual-process claim that the recollection process is always available is, in fact, more parsimonious than the single-process claim that the recollection process is used only in certain paradigms. The value of a well-specified process model such as the SAC model is discussed with regard to other types of dual-process models.

Deciding about decision models of remember and know judgments: A reply to Murdock (2006)

B. B. Murdock (2006) has interpreted remember-know data within a decision space defined by item and associative information, the fundamental variables in his general recognition memory model TODAM (B. B. Murdock, 1982). He has related parameters of this extended model to stimulus characteristics for several classic remember-know data sets. The authors show that this accomplishment is shared by both one- and two-dimensional signal-detection-based remember-know models (J. C. Dunn, 2004; C. M. Rotello, N. A. Macmillan, & J. A. Reeder, 2004; J. T. Wixted & V. Stretch, 2004), which can be represented in this same decision space and can be related to stimulus characteristics with similar success. Murdock claims that his model, unlike its competitors, is a process model; however, the process aspects of TODAM are not used in his application, and the decision aspects are identical to a previously proposed model. Murdock's claim that one competing model (STREAK; C. M. Rotello et al., 2004) is not fully specified is shown to be false. The new model is not superior to existing ones, but comparisons among the models to date are not definitive. The authors describe several strategies that might be applied to distinguish among them.

Memory strength and repetition suppression: multimodal imaging of medial temporal cortical contributions to recognition

Declarative memory permits an organism to recognize stimuli that have been previously encountered, discriminating them from those that are novel. One basis for recognition is item memory strength, which may support the perception of stimulus familiarity. Though the medial temporal lobes are known to be critical for declarative memory, at present the neural mechanisms supporting perceived differences in memory strength remain poorly specified. Here, functional MRI (fMRI) and anatomically constrained magnetoencephalography (MEG) indexed correlates of graded memory strength in the human brain, focusing on medial temporal cortex. fMRI revealed a decrease in medial temporal cortical activation that tracked parametric levels of perceived memory strength. Anatomically constrained MEG current estimates revealed that strength-dependent signal reductions onset within 150-300 ms. Memory strength appears to be rapidly signaled by medial temporal cortex through repetition suppression (activation reductions), providing a basis for the subjective perception of stimulus familiarity or novelty.

Theremember response: Subject to bias, graded, and not a process-pure indicator of recollection

Recognition memory judgments have long been assumed to depend on the contributions of two underlying processes: recollection and familiarity. We measured recollection with receiver-operating characteristic (ROC) data and remember-know judgments. Under standard remember-know instructions, the two estimates of recollection diverged. When subjects were told they might need to justify their remember responses to the experimenter, the two estimates were more likely to agree. The data support the conclusion that remember responses are generally based on a continuous underlying process but that specific task instructions can produce data that appear consistent with a high-threshold recollective process. Models based on signal detection theory provide a better account of these data than does the dual-process model (Yonelinas, 1994) or process-pure interpretations.

Parietal lobe contributions to episodic memory retrieval

Although the parietal lobe is not traditionally thought to support declarative memory, recent event-related fMRI studies of episodic retrieval have consistently revealed a range of memory-related influences on activation in lateral posterior parietal cortex (PPC) and precuneus extending into posterior cingulate and retrosplenial cortex. This article surveys the fMRI literature on PPC activation during remembering, a literature that complements earlier electroencephalography data. We consider these recent memory-related fMRI responses within the context of classical ideas about parietal function that emphasize space-based attention and motor intention. We conclude by proposing three hypotheses concerning how parietal cortex might contribute to memory.

fMRI evidence of word frequency and strength effects in recognition memory

We used event-related fMRI to investigate the neural correlates of encoding strength and word frequency effects in recognition memory. At test, participants made Old/New decisions to intermixed low (LF) and high frequency (HF) words that had been presented once or twice at study and to new, unstudied words. The Old/New effect for all hits vs. correctly rejected unstudied words was associated with differential activity in multiple cortical regions, including the anterior medial temporal lobe (MTL), hippocampus, left lateral parietal cortex and anterior left inferior prefrontal cortex (LIPC). Items repeated at study had superior hit rates (HR) compared to items presented once and were associated with reduced activity in the right anterior MTL. By contrast, other regions that had shown conventional Old/New effects did not demonstrate modulation according to memory strength. A mirror effect for word frequency was demonstrated, with the LF word HR advantage associated with increased activity in the left lateral temporal cortex. However, none of the regions that had demonstrated Old/New item retrieval effects showed modulation according to word frequency. These findings are interpreted as supporting single-process memory models proposing a unitary strength-like memory signal and models attributing the LF word HR advantage to the greater lexico-semantic context-noise associated with HF words due to their being experienced in many pre-experimental contexts.

On the importance of models in interpreting remember-know experiments: Comments on Gardiner et al.'s (2002) meta-analysis

From a meta-analysis of recognition experiments using the remember-know-guess paradigm, Gardiner, Ramponi, and Richardson-Klavehn (2002) reported two findings that they viewed as evidence against the one-dimensional model for that paradigm: (1) Memory strength increased when know responses were added to remember responses, decreasing when guess responses were also included. (2) The accuracy of guess responses was correlated with the location of the old-new criterion in the one-dimensional model for the paradigm, implying that guesses were influenced by decision processes. We question both findings. The first result is contradicted by a signal-detection (SDT) analysis, which shows that both know and guess responses reduced estimated memory strength. The discrepancy results from the properties of A', the measure of accuracy used by Gardiner et al., which we argue is flawed. The second result follows directly from the one-dimensional model, in which accuracy and response criteria are fixed. The authors' reasons for rejecting the one-dimensional model are thus not persuasive, but it can nonetheless be rejected because ROC curves implied by the data are inconsistent with ROCs derived from ratings experiments. A two-dimensional SDT model (Rotello, Macmillan, & Reeder, 2004) accounts for both sets of data. The analysis illustrates the importance of models in interpreting remember--know data.

Measuring the speed of the conscious components of recognition memory: remembering is faster than knowing

Three experiments investigated response times (RTs) for remember and know responses in recognition memory. RTs to remember responses were faster than RTs to know responses, regardless of whether the remember-know decision was preceded by an old/new decision (two-step procedure) or was made without a preceding old/new decision (one-step procedure). The finding of faster RTs for R responses was also found when remember-know decisions were made retrospectively. These findings are inconsistent with dual-process models of recognition memory, which predict that recollection is slower and more effortful than familiarity. Word frequency did not influence RTs, but remember responses were faster for words than for nonwords. We argue that the difference in RTs to remember and know responses reflects the time taken to make old/new decisions on the basis of the type of information activated at test.

The effects of word frequency and similarity on recognition judgments: the role of recollection

K. J. Malmberg, J. Holden, and R. M. Shiffrin (2004) reported more false alarms for low- than high-frequency words when the foils were similar to the targets. According to the source of activation confusion (SAC) model of memory, that pattern is based on recollection of an underspecified episodic trace rather than the error-prone familiarity process. The authors tested the SAC account by varying whether participants were warned about the nature of similar foils and whether the recognition test required the discrimination. More false alarms for low-frequency similar items occurred only when participants were not warned at study about the subtle features to be discriminated later. The differential false-alarm rate by word frequency corresponded to the pattern of remember responses obtained when the test instructions did not ask for a subtle discrimination, supporting the SAC account that reversed false-alarm rates to similar foils are based on the recollection process.

Recognition memory and introspective remember/know judgments: Evidence for the influence of distractor plausibility on “remembering” and a caution about purportedly nonparametric measures

One popular technique in the study of human recognition memory involves the elicitation of remember and know judgments and the attribution of those judgments to qualitative states of memory retrieval. An alternative view, reviewed here, implicates quantitative, but not qualitative, differences in evidence as the basis for those two judgments. That theory makes two clear and testable predictions: that of criterion shifts in "remembering" and that of isodiscriminability across different response sets. In this experiment, the makeup of the distractor set in a recognition test is shown to influence overall recognition criterion and also rates of "remember" responses. The second portion of the article demonstrates how A' is a poor choice of a measure to test the prediction of isodiscriminability. When this measure is corrected (Equation 7) to make it more consistent with current knowledge about the receiver-operating characteristic in recognition memory, it reveals that there is no difference in discriminability between "remember" and all positive responses.

Source-Monitoring Training: Toward Reducing Rater Expectancy Effects in Behavioral Measurement

The authors developed a source-monitoring procedure to reduce the biasing effects of rater expectations on behavioral measurement. Study participants (N = 224) were given positive or negative information regarding the performance of a group and, after observing the group, were assigned to a source-monitoring or control condition. Raters in the source-monitoring condition were instructed to report only behaviors that evoked detailed memories (remember judgments) and to avoid reporting behaviors based on feelings of familiarity (know judgments). Results revealed that controlling raters' response strategy reduced (and often eliminated) the biasing effects of performance expectations. These findings advance our understanding of the performance-cue bias and offer a potentially useful technique for decreasing rater bias.

Memory Awareness Following Speeded Compared with Unspeeded Picture Recognition

Abstract. A remember-know paradigm was used to assess memory awareness following speeded and unspeeded yes/no picture recognition. The beneficial effects of picture size congruency at study and test occurred with speeded as well as with unspeeded recognition. In each case, they were associated with remembering, not with knowing, which remained invariant. Thus, size congruency effects were associated with remembering even when recognition occurred more automatically and hence may be more dependent on a relatively fast familiarity process. In a second experiment, speeded remember responses were compared with remember responses that followed speeded yes/no recognition. There was more remembering when it was the remember responses that were speeded, contrary to what might be expected if remembering reflects a relatively slow recollection process. These results have implications for the ability of dual-process models of recognition memory to account for memory awareness.

Social Recognition Memory: The Effect of Other People's Responses for Previously Seen and Unseen Items

When people discuss their memories, what one person says can influence what another person reports. In 3 studies, participants were shown sets of stimuli and then given recognition memory tests to measure the effect of one person's response on another's. The 1st study (n = 24) used word recognition with participant-confederate pairs and found that the effect of confederate responses on participant responses was larger for previously unseen items than for previously seen items (omega(p) = .23). This finding was replicated in the 2nd study, which used photographs of cars (n = 24). In the 3rd study (n = 54), which used photographs of faces with participant pairs, the effect was also larger for unseen items. Results indicate that people rely more on other people's memories for unremembered objects than for remembered objects. This is important for both theories of memory and applications (e.g., witnesses talking, students studying together).

In defense of the signal detection interpretation of remember/know judgments

Donaldson (1996) argued that remember/know judgments can be conceptualized within a signal detection framework by assuming that they are based on two criteria situated along a strength-of-memory decision axis. According to this model, items that exceed a high criterion receive a remember response, whereas items that only exceed a lower criterion receive a know response. Although a variety of findings have been presented in evidence against this idea, Dunn (2004) recently showed that detection theory is fully compatible with those findings. We present a variety of new results and new analyses that weigh strongly in favor of the detection interpretation. We further show that a dual-process account of recognition memory is compatible with a unidimensional detection model despite the common notion that such a model necessarily assumes a single process. The key assumption of this model is that individual recognition decisions are based on both recollection and familiarity (not on one process or the other).

Recall and recognition are equally impaired in patients with selective hippocampal damage

In two recent studies, the hypothesis that recall is more severely impaired than recognition in patients with damage thought to be limited to the hippocampus was tested. Yonelinas et al. (2002) reported findings that appeared to support this hypothesis, whereas Manns, Hopkins, Reed, Kitchener, and Squire (2003) found that recall and recognition were equally impaired. An analysis of the individual subject data from the two studies revealed that the apparent disagreement stemmed from the inclusion of a single outlying recognition score for 1 of the 55 control subjects in Yonelinas et al. (2002). When that outlier was excluded, the studies were in agreement that recognition and recall are substantially and similarly impaired in these patients. Yonelinas et al. (2002) also analyzed remember/know judgments and ROC data in an effort to show that recollection is selectively impaired in patients, but these analyses also raise problems.

Sum-Difference Theory of Remembering and Knowing: A Two-Dimensional Signal-Detection Model

In the remember-know paradigm for studying recognition memory, participants distinguish items whose presentations are episodically remembered from those that are merely familiar. A one-dimensional model postulates that remember responses are just high-confidence old judgments, but a meta-analysis of 373 experiments shows that the receiver operating characteristic (ROC) curves predicted by this model have the wrong slope. According to the new sum-difference Theory of remembering and knowing (STREAK), old items differ from new ones in both global and specific memory strength: The old-new judgment is based on a weighted sum of these dimensions, and the remember- know judgment is based on a weighted difference. STREAK accounts for the form of several novel kinds of ROC curves and for existing remember-know and item-recognition data.