A signal detection theory analysis of an unconscious perception effect

The independent observation model (Macmillan & Creelman, 1991)is fitted to detection-identification data collected under conditions of heavy masking. The model accurately predicts a quantitative relationship between stimulus detection and stimulus identification over a wide range of detection performance. This model can also be used to offer a signal detection interpretation of the common finding of above-chance identification following a missed signal. While our finding is not a new one, the stimuli used in this experiment (redundant three-letter strings) differ slightly from those used in traditional signal detection work. Also, the stimuli were presented very briefly and heavily masked, conditions typical in the study of unconscious perception effects.     

Predicting similarity and categorization from identification.

In this article, the relation between the identification, similarity judgment, and categorization of multidimensional perceptual stimuli is studied. The theoretical analysis focused on general recognition theory (GRT), which is a multidimensional generalization of signal detection theory. In one application, 2 Ss first identified a set of confusable stimuli and then made judgments of their pairwise similarity. The second application was to Nosofsky's (1985b, 1986) identification-categorization experiment. In both applications, a GRT model accounted for the identification data better than Luce's (1963) biased-choice model. The identification results were then used to predict performance in the similarity judgment and categorization conditions. The GRT identification model accurately predicted the similarity judgments under the assumption that Ss allocated attention to the 2 stimulus dimensions differently in the 2 tasks. The categorization data were predicted successfully without appealing to the notion of selective attention. Instead, a simpler GRT model that emphasized the different decision rules used in identification and categorization was adequate.     

Stimulus intensity, attentional instructions, and the ear advantage during dichotic listening

The present experiment was designed to investigate the effects of the following two variables on the identification of dichotically presented CV syllables: (1) the relative difference in intensity levels of the stimuli presented to the two ears and (2) the instructions to attend to both ears or to focus attention on one ear. As expected for a verbal task, more CV's were identified from the right ear than from the left ear. Furthermore, identification of stimuli presented to one ear improved when (1) those stimuli were relatively higher in volume than the stimuli presented to the other ear and (2) when subjects were instructed to focus attention on only that ear rather than distribute attention across both ears. Of particular importance is the finding that the effects of relative stimulus intensity are the same under conditions of focused attention as under conditions of divided attention. This finding is inconsistent with an attention explantation of the relative intensity effects. Instead, the results are consistent with a model of dichotic listening in which ear of stimulus presentation and relative stimulus intensity influence a perceptual stage of information processing and attentional instructions influence a subsequent response selection stage.     

A comparison of signal detection theory to the objective threshold/strategic model of unconscious perception

A key problem in unconscious perception research is ruling out the possibility that weak conscious awareness of stimuli might explain the results. In the present study, signal detection theory was compared with the objective threshold/strategic model as explanations of results for detection and identification sensitivity in a commonly used unconscious perception task. In the task, 64 undergraduate participants detected and identified one of four briefly displayed, visually masked letters. Identification was significantly above baseline (i.e., proportion correct > .25) at the highest detection confidence rating. This result is most consistent with signal detection theory's continuum of sensory states and serves as a possible index of conscious perception. However, there was limited support for the other model in the form of a predicted "looker's inhibition" effect, which produced identification performance that was significantly below baseline. One additional result, an interaction between the target stimulus and type of mask, raised concerns for the generality of unconscious perception effects.     

Rapid acquisition of bias in signal detection: dynamics of effective reinforcement allocation

We investigated changes in bias (preference for one response alternative) in signal detection when relative reinforcer frequency for correct responses varied across sessions. In Experiment 1, 4 rats responded in a two-stimulus, two-response identification procedure employing temporal stimuli (short vs. long houselight presentations). Relative reinforcer frequency varied according to a 31-step pseudorandom binary sequence and stimulus duration difference varied over two values across conditions. In Experiment 2, 3 rats responded in a five-stimulus, two-response classification procedure employing temporal stimuli. Relative reinforcer frequency was varied according to a 36-step pseudorandom ternary sequence. Results of both experiments were analyzed according to a behavioral model of detection. The model was extended to incorporate the effects of current and previous session reinforcer frequency ratios on current-session performance. Similar to findings with concurrent schedules, effects on bias of relative reinforcer frequency were highest for the current session. However, carryover from reinforcer ratios of previous sessions was evident. Generally, the results indicate that bias can come under control of frequent changes in relative reinforcer frequency in both identification and classification procedures.     

The object-detection effect: configuration enhances perception.

Line drawings used by Weisstein and Harris (1974) are seen as box-like three-dimensional figures if the lines are arranged properly. A flat two-dimensional pattern is seen when these same lines are disarranged. A target line contained within the three-dimensional figure is identified more readily than is the same line contained within a two-dimensional figure. This finding was extended in the present experiments: The three-dimensional stimulus was detected more quickly than the two-dimensional stimulus, under conditions of visual backward masking. Three-dimensional stimuli were also classified more quickly than two-dimensional stimuli. Just as with the face-detection effect and the word-detection effect, object detection can be affected by the form of the visual stimulus.     

Feature sampling in detection: implications for the measurement of perceptual independence

The article presents the feature sampling signal detection (FS-SDT) model, an extension of the multivariate signal detection (SDT) model. The FS-SDT model assumes that, because of attentional shifts, different subsets of features are sampled for different presentations of the same multidimensional stimulus. Contrary to the SDT model, the FS-SDT model enables the estimation of pure perceptual effects that are uncontaminated by strategic attention shifts. The consideration of feature sampling in detection and identification opens a new perspective on the problem of measuring, respectively, the separability and integrality of stimulus dimensions. Disregarding feature sampling as a component process in detection and identification usually results in biased estimations of perceptual independence concepts relevant for judgments of whether stimulus dimensions are processed independently.     

The object-detection effect: Configuration enhances perception

Line drawings used by Weisstein and Harris (1974) are seen as box-like three-dimensional figures if the lines are arranged properly. A flat two-dimensional pattern is seen when these same lines are disarranged. A target line contained within the three-dimensional figure is identified more readily than is the same line contained within a two-dimensional figure. This finding was extended in the present experiments: The three-dimensional stimulus was detected more quickly than the two-dimensional stimulus, under conditions of visual backward masking. Three-dimensional stimuli were also classified more quickly than two-dimensional stimuli. Just as with the face-detection effect and the word-detection effect, object detection can be affected by the form of the visual stimulus.     

Contiguity and conditioned reinforcement in probabilistic choice

In a baseline condition, pigeons chose between an alternative that always provided food following a 30-s delay (100% reinforcement) and an alternative that provided food half of the time and blackout half of the time following 30-s delays (50% reinforcement). The different outcomes were signaled by different-colored keylights. On average, each alternative was chosen approximately equally often, replicating the finding of suboptimal choice in probabilistic reinforcement procedures. The efficacy of the delay stimuli (keylights) as conditioned reinforcers was assessed in other conditions by interposing a 5-s gap (keylights darkened) between the choice response and one or more of the delay stimuli. The strength of conditioned reinforcement was measured by the decrease in choice of an alternative when the alternative contained a gap. Preference for the 50% alternative decreased in conditions in which the gap preceded either all delay stimuli, both delay stimuli for the 50% alternative, or the food stimulus for the 50% alternative, but preference was not consistently affected in conditions in which the gap preceded only the 100% delay stimulus or the blackout stimulus for the 50% alternative. These results support the notion that conditioned reinforcement underlies the finding of suboptimal preference in probabilistic reinforcement procedures, and that the signal for food on the 50% reinforcement alternative functions as a stronger conditioned reinforcer than the signal for food on the 100% reinforcement alternative. In addition, the results fail to provide evidence that the signal for blackout functions as a conditioned punisher.     

Detecting and identifying change: additions versus deletions

The recognition of changes in the features of objects was examined as a function of the nature of the change (additions, deletions, no change). In two experiments we assessed both detection (noticing whether the stimulus had changed) and identification (specifying the exact nature of the change). Both detection and identification were expected to depend upon the subject's awareness of the subsequent recognition tasks while the original stimuli were encoded. In Experiment 1, subjects were not aware of subsequent detection and identification tasks while they initially viewed study slides of the to-be-changed stimuli. During the subsequent presentation of the test stimuli, detection and identification were superior for additions. On the other hand, in Experiment 2 when subjects were aware of the subsequent recognition tasks while viewing each study slide, a detection advantage for deletions obtained. Identification performance depended upon a further factor, whether the features of a stimulus were codable. Only in codable stimuli were deletions easier to identify than additions. The differences between the two experiments in detecting and identifying additions versus deletions are consistent with Tversky's (1977) research that stresses the importance of specifying which representation (the study stimulus or the test stimulus) is the subject of comparison in the comparative judgment.     

Unconscious learning. Conditioning to subliminal visual stimuli

The role of consciousness in Pavlovian conditioning was examined in two experiments in which visually masked neutral words were used as the conditioned stimuli (CS) and an electric shock as the unconditioned stimulus (US). The inter-stimulus interval (ISI) was established individually. A detection threshold was used in Experiment 1 and an identification threshold in Experiment 2. The primary dependent variable was the skin conductance response (SCR). Results showed that the conditioned response (CR) was acquired by 58% of participants who perceived stimuli above the identification threshold, 50% of participants who perceived stimuli below the detection threshold, and 11% of participants who perceived stimuli below the identification threshold, but above the detection threshold. These results suggest that consciousness of the CS-US contingency is not a necessary condition for acquiring a CR of the autonomous nervous system (ANS).     

Effects of hypoxia on peripheral visual response to dim stimuli.

Response times (RTs) of 9 Ss were obtained for detection of 48 flash stimuli distributed throughout the visual field during 3 1/4-hr. exposures to each of 4 hypoxia conditions (0, 13,000, 15,000, 17,000 feet equivalent elevation). The luminances of all stimuli were set in common at the detection threshold value for the visual periphery. RTs were impaired in direct relation to hypoxic exposure severity, the peak impairments occurring within 90 min. followed by gradual recovery. Since the present results showed less impairment than previous data for brighter stimuli using the same task, it is concluded that stimulus contrast is more critical to peripheral signal detection than absolute stimulus luminance, particularly under hypoxic exposure.     

Confidence in word detection predicts word identification: implications for an unconscious perception paradigm

The present experiments extend the scope of the independent observation model based on signal detection theory (Macmillan & Creelman, 1991) to complex (word) stimulus sets. In the first experiment, the model predicts the relationship between uncertain detection and subsequent correct identification, thereby providing an alternative interpretation to a phenomenon often described as unconscious perception. Our second experiment used an exclusion task (Jacoby, Toth, & Yonelinas, 1993), which, according to theories of unconscious perception, should show qualitative differences in performance based on stimulus detection accuracy and provide a relative measure of conscious versus unconscious influences (Merikle, Joordens, & Stoltz, 1995). Exclusion performance was also explained by the model, suggesting that undetected words did not unconsciously influence identification responses.     

Effects of relative reinforcer frequency on complex color detection

Pigeons were trained under a discrete-trials detection procedure in which one of a set of color stimuli was presented on the center key and a single response turned off the stimulus and illuminated two side keys. Single responses to one or the other side key produced occasional reinforcers depending on the value of the color stimulus. In Experiment 1, one color-stimulus set comprised 559, 564, 569, and 574 nm, and right-key pecks were occasionally reinforced following presentations of members of this set. The other stimulus set comprised 579, 584, 589, and 594 nm, and left-key pecks were occasionally reinforced following presentations of members of this set. Across seven experimental conditions, the left/(left + right) relative reinforcer frequency was varied from .1 to .9. In Experiment 2, one stimulus set contained only one member, 574 nm, and right-key responses were occasionally reinforced following its presentation. Over 12 experimental conditions, two manipulations were carried out. First, the number of stimuli comprising the other stimulus set was increased from one (579 nm) to two (579 and 584 nm) to three (579, 584, and 589 nm) and to four (579, 584, 589, and 594 nm), and left-key responses were reinforced occasionally following center-key presentations of members of this set. Second, for each stimulus combination, the left/(left + right) relative reinforcer frequency was varied from .1 to .5 to .9 across three experimental conditions. The principal finding of Experiments 1 and 2 was that reinforcers and stimuli interacted in their effects on behavior. In Experiment 3, pairs of adjacent stimuli (5 nm apart) in the range 559 to 594 nm were presented in each experimental condition, and the left/(left + right) relative reinforcer frequency was held constant at .5. The data from all three experiments were analyzed according to a detection model describing performance in multiple-stimulus two-response procedures. This model provided independent measures of stimulus discriminability, contingency discriminability, and bias. The analysis showed that (a) consistent with the color-naming function, pigeons were better able to discriminate between higher nanometer values than lower nanometer values; (b) their ability to discriminate between the stimuli was independent of the number of wavelengths comprising each stimulus set; (c) they allocated delivered reinforcers very accurately to the previously emitted response; and (d) no consistent biases emerged.     

Electrophysiological evidence for a shared representational medium for visual images and visual percepts

Does mental imagery involve the activation of representations in the visual system? Systematic effects of imagery on visual signal detection performance have been used to argue that imagery and the perceptual processing of stimuli interact at some common locus of activity (Farah, 1985). However, such a result is neutral with respect to the question of whether the interaction occurs during modality-specific visual processing of the stimulus. If imagery affects stimulus processing at early, modality-specific stages of stimulus representation, this implies that the shared stimulus representations are visual, whereas if imagery affects stimulus processing only at later, amodal stages of stimulus representation, this implies that imagery involves more abstract, postvisual stimulus representations. To distinguish between these two possibilities, we repeated the earlier imagery-perception interaction experiment while recording event-related potentials (ERPs) to stimuli from 16 scalp electrodes. By observing the time course and scalp distribution of the effect of imagery on the ERP to stimuli, we can put constraints on the locus of the shared representations for imagery and perception. An effect of imagery was seen within 200 ms following stimulus presentation, at the latency of the first negative component of the visual ERP, localized at the occipital and posterior temporal regions of the scalp, that is, directly over visual cortex. This finding provides support for the claim that mental images interact with percepts in the visual system proper and hence that mental images are themselves visual representations.     

Visual awareness and the detection of fearful faces

A commonly held view is that emotional stimuli are processed independently of awareness. Here, the authors parametrically varied the duration of a fearful face target stimulus that was backward masked by a neutral face. The authors evaluated awareness by characterizing behavioral performance using receiver operating characteristic curves from signal detection theory. Their main finding was that no universal objective awareness threshold exists for fear perception. Although several subjects displayed a behavioral pattern consistent with previous reports (i.e., targets masked at 33 ms), a considerable percentage of their subjects (64%) were capable of reliably detecting 33-ms targets. Their findings suggest that considerable information is available even in briefly presented stimuli (possibly as short as 17 ms) to support masked fear detection.     

Restoring the associability of a pre-exposed CS by a surprising event

Three experiments are reported which use rats and the conditioned suppression technique. The first two confirmed a previous finding that prior exposure to a stimulus predicting a weak shock retards further learning when this same stimulus is subsequently used to signal a stronger shock. They further showed that this loss of stimulus associability could be attenuated by inserting trials on which the stimulus was presented alone in the absence of shock before the phase of training with the stronger shock. Experiment III demonstrated that, for animals given prior exposure to two stimuli, the insertion of nonreinforced trials with one of the stimuli will restore the associability only of that stimulus. These results are taken to show that a surprising event (the omission of an expected shock) can restore the associability of a pre-exposed conditioned stimulus.     

The effects of stimulus dimensionality on the rate of gain of information

Hick's paradigm as extended to an absolute judgment task by Doherty was used to study the effects of dimensionality on the rate of stimulus identification in two experiments. Two undimensional conditions, size and brightness, and two bidimensional conditions, size-brightness redundant and size-brightness non-redundant, were employed. The significant linear components of the regression of choice reaction time on transmitted information for the undimensional and bidimensional stimuli supported and extended Hick's law. The finding of nearly identical slopes for the two undimensional and redundant bidimensional condition regression lines suggests a stage of processing which has a constant capacity in bits/s and which is independent of stimulus dimensionality. An increase in slope for the non-redundant condition is attributed to a difference in response requirements.     

Divided attention between simultaneous auditory and visual signals

Past studies of simultaneous attention to pairs of visual stimuli have used the “dual-task” paradigm to show that identification of the direction of a change in luminance, whether incremental or decremental, is “capacity-limited,” while simple detection of these changes is governed by “capacity-free” processes. On the basis of that finding, it has been suggested that the contrast between identification and detection reflects different processes in the sensory periphery, namely the responses of magno- and parvocellular receptors. The present study questions that assertion and investigates the contribution of central processing in resource limitation by applying the dual task to a situation in which one stimulus is auditory and one is visual. The results are much the same as before, with identification demonstrating the tradeoff in performance generally attributed to a limited capacity but detection showing no loss compared with single-task controls. This implies that limitations on resources operate at a central level of processing rather than in the auditory and visual peripheries.     

Differential redundancy gain in onset detection versus offset detection

In a visual simple reaction time paradigm with attention divided between the left and right visual fields, redundancy gain refers to the finding of faster responses to stimuli presented in both fields than to single stimuli. The present study investigated whether the effects of low-level perceptual processing affect the redundancy gain by comparing the detection of onsets versus offsets. In different blocks, participants responded to left and right visual field stimuli that either appeared (onset) or disappeared (offset), with a single stimulus change in some trials and two redundant stimulus changes in others. With onset stimuli, the results replicated previous redundancy gain effects. In contrast, there was much less redundancy gain when participants responded to stimulus offsets. This finding suggests that redundancy gain is sensitive to low-level perceptual characteristics, such as onset versus offset presentation of stimuli.