Recognition Confidence Research Articles

Patients with schizophrenia do not produce more false memories than controls but are more confident in them

Patients diagnosed with schizophrenia consistently demonstrate impairment in memory acquisition. However, no empirical consensus has been achieved on whether or not patients are more prone to produce false memories. A visual variant of the Deese-Roediger-McDermott (DRM) paradigm was administered to 35 schizophrenia patients and 34 healthy controls. Recognition and recognition confidence were later tested for studied and lure items. Strong contextual cues at recognition encouraged adoption of a gist-based retrieval strategy, which was predicted to elicit over-confidence in errors and increase the false memory rate in patients. Patients were significantly impaired on true item recognition but did not display more false memories than healthy subjects. As predicted from prior findings by our group, patients were more confident than controls for lure items, while being at the same time under-confident for studied items (reduced confidence gap). Although patients did not produce more false memories than controls, such errors were made with higher confidence relative to controls. The decreased confidence gap in patients is thought to stem from a gist-based recollection strategy, whereby little evidence suffices to make a strong judgment.

Correlating call data and speech recognition information in a telephony applicaiton

A method for improving the recognition confidence of alphanumeric spoken input, suitable for use in a speech recognition telephony application such as a voice response system. An alphanumeric candidate is determined from the spoken input, which may be the best available representation of the spoken input. Recognition confidence is compared with a preestablished threshold. If the recognition confidence exceeds the threshold, the alphanumeric candidate is selected to represent the spoken input. Otherwise, present call data associated with the spoken input is determined. Call data may include automatic number identification (ANI) information, caller-ID information, and/or dialed number information service (DNIS) information. Information associated with the alphanumeric candidate and information associated with the present call data are correlated in order to select alphanumeric information that best represents the spoken input.

Evaluation of an online bioterrorism continuing medical education course

Much of the international community has an increased awareness of potential biologic, chemical, and nuclear threats and the need for physicians to rapidly acquire new knowledge and skills in order to protect the public's health. The present study evaluated the educational effectiveness of an online bioterrorism continuing medical education (CME) activity designed to address clinical issues involving suspected bioterrorism and reporting procedures in the United States. This was a retrospective survey of physicians who had completed an online CME activity on bioterrorism compared with a nonparticipant group who had completed at least 1 unrelated online CME course from the same medical school Web site and were matched on similar characteristics. An online survey instrument was developed to assess clinical and systems knowledge and confidence in recognition of illnesses associated with a potential bioterrorism attack. A power calculation indicated that a sample size of 100 (50 in each group) would achieve 90% power to detect a 10% to 15% difference in test scores between the two groups. Compared with nonparticipant physicians, participants correctly diagnosed anthrax (p = .01) and viral exanthem (p = .01), but not smallpox, more frequently than nonparticipants. Participants knew more frequently than nonparticipants who to contact regarding a potential bioterrorism event (p = .03) Participants were more confident than nonparticipants about finding information to guide diagnoses of patients with biologic exposure (p = .01), chemical exposure (p = .02), and radiation exposure (p = .04). An online bioterrorism course shows promise as an educational intervention in preparing physicians to better diagnose emerging rare infections, including those that may be associated with a bioterrorist event, in increasing confidence in diagnosing these infections, and in reporting of such infections for practicing physicians.

The relation between realism in confidence judgements and the phenomenological quality of recognition memory when using emotionally valenced pictures

The aim of the study was to investigate the relation between the phenomenological quality of memory (Tulving, 1985) and the realism (validity) in confidence judgement when using emotional pictures (I.A.P.S; Lang, Ohman, & Vaitl, 1988). A series of three experiments was completed where the participants judged the phenomenological quality of their memory and/or their confidence. The results showed facilitation for the negative pictures in a matrix search task in the encoding phase, where negative pictures were more easily and quickly detected, compared to positive ones. In the memory phase of the experiments a higher degree of recollective experience (a larger proportion of 'remember' responses) was found for negative pictures. A higher level of confidence for recognition of negative pictures than for positive ones was obtained, but no general valence dependent effect on the realism in the confidence judgement was found. However, when analysing only the remember responses, negative pictures showed higher overconfidence than the positive pictures. The results support that a recollective experience induces higher confidence and overconfidence. Copyright (c) 2006 John Wiley & Sons, Ltd. (Less)

Postidentification feedback effects on face recognition confidence: evidence for metacognitive influences

Information given to witnesses after an identification decision greatly alters their impressions of the original event and importantly, their identification confidence. Two experiments investigated the possibility that the effect of feedback on confidence may be altered according to the strength of the witness's cues to accuracy. Experiment 1 used a manipulation of exposure duration to alter recognition accuracy prior to the delivery of confirming, disconfirming or no feedback. While the feedback effect was not different across exposure duration conditions, decisions that were made more quickly were less likely to show large changes in confidence due to feedback. Experiment 2 manipulated the distinctiveness of faces and showed that the effects of feedback on confidence, and on the resolution of the confidence judgement, were more pronounced when disconfirming feedback was given for distinctive faces and when confirming feedback was given for typical faces. These studies showed that the impressions that participants formed of their likely accuracy might moderate the effects of feedback on decision confidence. Copyright © 2006 John Wiley & Sons, Ltd.

Effects of healthy aging on hippocampal and rhinal memory functions: an event-related fMRI study.

Event-related functional magnetic resonance imaging was used to study the effects of healthy aging on hippocampal and rhinal memory functions. Memory for past events can be based on retrieval accompanied by specific contextual details (recollection) or on the feeling that an event is old or new without the recovery of contextual details (familiarity). There is evidence that recollection is more dependent on hippocampus, whereas familiarity is more dependent on the rhinal cortex, and that healthy aging has greater effects on recollection than on familiarity. However, little evidence is available about the neural correlates of these effects. Here, we isolated activity associated with recollection and familiarity by distinguishing between linear and quasi-exponential "perceived oldness" functions derived from recognition confidence levels. The main finding was a double dissociation within the medial temporal lobes between recollection-related activity in hippocampus, which was reduced by aging, and familiarity-related activity in rhinal cortex, which was increased by aging. In addition, age dissociations were found within parietal and posterior midline regions. Finally, aging reduced functional connectivity within a hippocampal-retrosplenial/parietotemporal network but increased connectivity within a rhinal-frontal network. These findings indicate that older adults compensate for hippocampal deficits by relying more on rhinal cortex, possibly through a top-down frontal modulation. This finding has important clinical implications because early Alzheimer's disease impairs both hippocampus and rhinal cortex.

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.

Integration of speech recognition and machine translation: Speech recognition word lattice translation

An important issue in speech translation is to minimize the negative effect of speech recognition errors on machine translation. We propose a novel statistical machine translation decoding algorithm for speech translation to improve speech translation quality. The algorithm can translate the speech recognition word lattice, where more hypotheses are utilized to bypass the misrecognized single-best hypothesis. The decoding involves converting the recognition word lattice to a translation word graph by a graph-based search, followed by a fine rescoring by an A* search. We show that a speech recognition confidence measure implemented by posterior probability is effective to improve speech translation. The proposed techniques were tested in a Japanese-to-English speech translation task, in which we measured the translation results in terms of a number of automatic evaluation metrics. The experimental results demonstrate a consistent and significant improvement in speech translation achieved by the proposed techniques.

Awareness of Heart Attack Symptoms and Lifesaving Actions Among New York City Area Residents

The American Heart Association has a national network of community-based programs designed to reduce response times to cardiac emergencies by improving access to automatic external defibrillators (AEDs) among laypersons. Success of these Operation Heartbeat programs depends in part on the public's knowledge of the warning signs of a myocardial infarction (MI) and appropriate response to cardiac arrest victims. In May 2000, a 7-minute telephone survey was administered to a random sample of adults residing within the American Heart Association affiliate territories of New York, New Jersey, and Connecticut to determine the knowledge of MI symptoms, confidence in cardiopulmonary resuscitation (CPR) use, and the awareness of AEDs. Of the respondents, 60% were women (n=1,128), 83% were Caucasians (n=1,558), 15.2% were non-whites (African American, Asian, or Hispanic), and 38.5% had at least a college degree (n=724). Women were significantly more likely than men to know that sex differences exist in the warning signs for an MI (63% vs. 30.7%, respectively; P< .001). Whites had above-average confidence in MI recognition compared with non-whites (39.2% vs. 27.4%, respectively; P< .001) and were more cognizant of the public availability of AEDs (54.5% vs. 33.2%, respectively; P< .001). Our findings suggest that racial/ethnic and sex disparities exist in the awareness of AEDs and in the knowledge of atypical MI symptoms in women, respectively. Innovative CPR outreach programs might be needed in New York area communities to increase CPR training among all adults, to increase AED awareness in vulnerable populations, and to improve knowledge and confidence in the recognition of acute MI symptoms.

Separating the Brain Regions Involved in Recollection and Familiarity in Recognition Memory

The neural substrates of recognition memory retrieval were examined in a functional magnetic resonance imaging study designed to separate activity related to recollection from that related to continuous variations in familiarity. Across a variety of brain regions, the neural signature of recollection was found to be distinct from familiarity, demonstrating that recollection cannot be attributed to familiarity strength. In the prefrontal cortex, an anterior medial region was related to recollection, but lateral regions, including the anterior and dorsolateral prefrontal cortex, were related to familiarity. Along the lateral parietal cortex, two functionally distinct regions were also observed: a lateral parietal/temporal region related to recollection and a more superior parietal region involved in familiarity. Similarly, in medial parietal regions, the posterior cingulate was related to recollection, whereas the precuneus was related to familiarity. The hippocampus was related to recollection, but also exhibited an inverse relationship to familiarity-driven recognition confidence. The results indicate that recollection and familiarity rely on different networks of brain regions and provide insights into the functional roles of different regions involved in episodic recognition memory.

Transient Disruption of Ventrolateral Prefrontal Cortex During Verbal Encoding Affects Subsequent Memory Performance

Episodic memory supports conscious remembrance of everyday experience. Prior functional neuroimaging data indicate that episodic encoding during phonological task performance is correlated with activation in bilateral posterior ventrolateral prefrontal cortex (pVLPFC), although uncertainty remains regarding whether these prefrontal regions make necessary contributions to episodic memory formation. Using functional MRI data to guide application of single-pulse transcranial magnetic stimulation (spTMS), this study examined the necessity of left and right pVLPFC for episodic encoding (as expressed through subsequent memory performance). To assess the timing of critical computations, pVLPFC function was transiently disrupted at different poststimulus onset times while subjects made syllable decisions about visually presented familiar and unfamiliar words; subsequent memory for these stimuli was measured. Results revealed that left pVLPFC disruption during encoding of familiar words impaired subsequent memory, expressed as a decline in recognition confidence, with disruption being maximal at 380 ms after stimulus onset. In contrast, right pVLPFC disruption facilitated subsequent memory for familiar words, expressed as an increase in medium confidence recognition, with this facilitation being maximal at 380 ms. Finally, phonological (syllable) decision accuracy was facilitated by right pVLPFC disruption, with this effect being maximal at 340 ms, but was unaffected by left pVLPFC disruption. These findings suggest that left pVLPFC mechanisms onset between 300 and 400 ms during phonological processing of words, with these mechanisms appearing necessary for effective episodic encoding. In contrast, disruption of correlated mechanisms in right pVLPFC facilitates encoding, perhaps by inducing a functional shift in the mechanisms engaged during learning.

Post-dialogue confidence scoring for unsupervised statistical language model training

This paper presents a new recognition confidence scoring method for unsupervised training of statistical language models in spoken language dialogue systems. Based on the proposed confidence scoring, the speech recognition results for untranscribed user utterances are selected for training the statistical language models of speech recognizers. The method uses features that can only be obtained after the dialogue session, in addition to other features, such as the acoustic scores of recognition results. Experimental results show that the proposed confidence scoring improves correct/incorrect classification of recognition results and that using the language models obtained through our approach results in better recognition accuracy than that achieved by conventional methods.

Knowledge-Combining Methodology for Dialogue Design in Spoken Language Systems

In this paper, we propose a strategy for designing dialogue managers in spoken dialogue systems for a restricted domain. This strategy combines several information sources intuition, observation and simulation, in order to maximize the adaptation within the system capability and the expectation of the user. These sources are combined by an iterative process consisting of five steps, where different dialogue alternatives are proposed and evaluated sequentially. The evaluation process includes different measures depending on the information required. Several measures are proposed and analyzed in each step. We also describe a user-modeling technique and an approach for designing the confirmation sub-dialogues based on recognition confidence measures. The knowledge-combining methodology is described and applied to a railway information system. In a subjective evaluation, users from the university gave the system a 3.9 score on a 5-point scale with an average call duration of 205 seconds. The employers of the railway company were more critical of the system. They gave it a score of 2.1 even though the system resolved more than half of the calls (57.8%) within an average call duration of three minutes (185 seconds).

The Claim Tool Kit for ad hoc recognition of peer entities

In ubiquitous/pervasive computing environments, it is envisaged that computing elements—entities—will start interacting in an ad hoc fashion. The peer-to-peer (p2p) paradigm is appealing for such types of interaction especially with JXTA, which supports the development of reusable p2p building blocks, which facilitate implementation on any smart device. However, the inability to rely on a centralised authentication infrastructure, the openness of the environment and the absence of an administrator (it is assumed to be too expensive to have a skilled administrator at hand due to the large number of peers) challenge the use of legacy authentication mechanisms. Supporting spontaneous interactions among previously unknown entities requires dynamic enrolment of strangers and unknown entities. Entity recognition (ER) is a process that is carried out each time an interaction happens between entities in order to dynamically recognise previously met entities. In this paper, we present the Claim Tool Kit (CTK), a Java-based implementation of ER: entities exchange messages, called Claims, and rely on their associated clues to evaluate the level of confidence in recognition. The CTK employs advanced features available with Java, such as JXTA and Java Cryptography and Security Architectures. We show that the CTK needs performance results on these features in order to increase the level of auto-configuration of the CTK. We describe how to obtain performance assessment for some of these new features. Finally, we explain how the CTK can be instrumented to take into account performance assessment. By analysing the evaluation results, the applicability of these advanced Java-based technologies for peer entity recognition is assessed.

The SENECA spoken language dialogue system

This article describes a speech-based user interface to a wide range of entertainment, navigation and communication applications in mobile environments by means of human-machine dialogues. The system has been developed in the framework of the EU-project SENECA. It uses noise reduction, speech recognition, and dialogue processing techniques. One interesting aspect relies in the fact that low speech recognition confidence and word-level ambiguities are compensated by engaging flexible clarification dialogues with the user. The SENECA system demonstrator has been evaluated by means of user tests. With speech input, road safety, especially for complex tasks is significantly improved. Compared to manual input, the feeling of being distracted from driving is less important with speech.

Judgment of frequency versus recognition confidence: repetition and recursive reminding.

Judgments of presentation frequency (JOFs) were compared with recognition confidence ratings (RCRs) in a single memory experiment. Two differences were found: (1) Relative to the effect of exposure duration, frequency had a larger effect on JOF than it had on RCR. (2) Replicating a finding by Proctor (1977), normalized memory operating characteristic (zMOC) curves for JOF had slopes greater than 1.0, whereas those for RCR had slopes of less than 1.0. The slope difference was found to be attributable to the first study trial. The results are contrary to the hypothesis that a single strength or familiarity dimension underlies JOF and RCR. To explain both findings, a new hypothetical basis of JOF is proposed. Repetition is assumed to trigger study phase reminding, which, in turn, is encoded into memory. Remindings can be recursively embedded, and the depth of recursion, recollected at test, is the primary basis of JOF. The hypothesis appears consistent with a broad range of JOF findings.

Psychopathology and illness beliefs influence COPD self-management

Objective To explore the influence of psychological characteristics in Chronic Obstructive Pulmonary Disease (COPD) self-management. Methods Patients admitted with an exacerbation of COPD were interviewed for psychiatric symptoms, illness beliefs and self-management behaviour using a new COPD Self-Management Interview (COPD-SMI). This comprised three scenarios to mimic a future evolving exacerbation. Responses were scored for knowledge and actions (adherence) for each scenario. Results Of 47 people approached, 39 participated; 41% had panic attacks, 33% general anxiety, 35% a depression history, 31% an anxiety history and 21% an alcohol dependence history. Twenty-six (67%) had a self-management plan. When hypothetically “well” lower (poorer) COPD-SMI Knowledge Scores were associated with an alcohol dependence history ( P=.025), no panic ( P=.021) and males ( P=.028). Those perceiving less influence over COPD had lower Action Scores during this scenario ( P=.01) and the “early exacerbation” scenario ( P=.05). Lower Knowledge Scores for the “early exacerbation” were associated with no panic ( P=.01) and no self-management plans ( P=.03). For the “severe exacerbation”, lower Action Scores were associated with depression history ( P=.004), panic ( P=.002), higher FEV 1% and no self-management plans ( P=.005). Higher PaCO 2 was associated with lower confidence in symptom recognition, self-management ability and medical care influencing COPD. Conclusion Anxiety, depression, alcohol use and illness beliefs may differentially influence self-management. Depression, previous alcohol dependence and perceived less influence over COPD inhibited self-management. Those with panic demonstrated more self-management knowledge when “well” but performed poorly on actions during the “severe exacerbation”. Those with self-management plans had better knowledge and actions.

False memories in schizophrenia.

In prior studies, it was observed that patients with schizophrenia show abnormally high knowledge corruption (i.e., high-confident errors expressed as a percentage of all high-confident responses were increased for schizophrenic patients relative to controls). The authors examined the conditions under which excessive knowledge corruption occurred using the Deese-Roediger-McDermott paradigm. Whereas knowledge corruption in schizophrenia was significantly greater for false-negative errors relative to controls, no group difference occurred for false-positive errors. The groups showed a comparable high degree of confidence for false-positive recognition of critical lure items. Similar to findings collected in elderly participants, patients, but not controls, showed a strong positive correlation between the number of recognized studied items and false-positive recognition of the critical lure.

Knowing What You Cannot Recognise: Further Evidence for Intact Metacognition in Alzheimer's Disease

Previous research has examined whether Alzheimer's disease (AD) patients can proficiently monitor their memory processing. In the present study we extended the scope of this research to consider confidence in recognition in an episodic memory task, using a forced-choice recognition paradigm to avoid floor effects. AD patients and age-matched controls studied 32 words, followed immediately by a recognition test requiring participants to select the target from pairs of items. Participants rated confidence in their choice on a 3-point scale. Despite significantly poorer memory performance and less confidence overall, gamma correlations indicated that the AD group accurately assigned confidence judgements to their responses. This adds support to the claim that metacognitive monitoring remains intact in AD.

Judgments of recency and their relation to recognition memory.

Subjects went through a list of 550 high- and low-frequency words (Experiment 1) or concrete and abstract words (Experiment 2) in which individual items were repeated at lags of 5 to 30 other items. They made old versus new recognition decisions on each word and followed each "old" response with a numerical judgment of recency (JOR). Recognition judgments displayed the mirror effect. Conditionalized on recognition, JORs were shorter for low-frequency words than for high-frequency words, and shorter for concrete words than for abstract words. This was true at every lag, suggesting that recognition and JOR may have a common basis. However, recognition confidence ratings obtained in Experiment 3 proved much less sensitive than JOR to test lag. Memory models applicable to multiple judgment tasks will be needed to account for such findings.