Harold Pashler - US grants

When people attempt to perform two simple sensorimotor tasks close together in time, performance deteriorates. Detailed chronometric analysis of this dual-task inerference can provide powerful empirical constraints on theodes of the mechanisms that carry out perceptual, cognitive and response processes. in previous work, the Pi has found converging evidence for a theory of dual-task interference that postulates two distinct components. When complex visual stimuli are presented at the same time, perceptual processing proceeds simultaneously, with impairments in accuracy dependent upon complexity. When response selection is required on one task, however, response selection on another task must wait: this processing stage is subject to discrete queueing. The work proposed here will address four questions, to test, elaborate and extend this theory. First, perceptual interference will be examined in auditory discrmination tasks (combined with other auditory or visual tasks), to determine whether perceptual interference reflects separate pools of resources tied to particular sensory modalities. Second, several different methods will be employed to determine whether post-perceptual processes other than response selection are subject to discrete queueing; this work will focus on semantic categorization and shifts of visual attention. Third, the possibility of independent control of different response modalities will be examined, focussing on the case of oculomotor control (saccadic and smooth pursuit eye movements). Pilot data suggests a surprising degree of independence in certain oculomotor tasks. Finally, we will investigate how extensive practice reduces dual-task interference. Previous work has suggested that well-practiced tasks may become completely "automatized", but the evidence is open to various interpretations; chronometric methods will be used to determine exactly which components of interference are affected by practice, and how they are affected.

Human beings are subjected to a vast barrage of sensory inputs. Typically, however, we attend to only a tiny subset of these inputs; our awareness and later memory of sensory events is mostly confined to this subset. While research has revealed something of how the brain's selection processes works, this knowledge is largely confined to artificial tasks in which people are explicitly instructed to attend to arbitrarily defined classes of stimuli. The present studies will explore mechanisms and principles that govern the spontaneous allocation of visual attention in the absence of such specific task demands, or in some cases, even work in opposition to voluntary task goals. One set of studies will examine the finding that visual inputs related to the contents of short-term memory automatically receive attentional priority, even when the individual attempts to prevent this. Other studies will examine the commonly observed tendency for an individual's own name to draw his or her attention; the limitations and causes of this phenomenon will be explored. A third set of studies will examine the effects of stimulus novelty, while the final set will examine the tendency of emotionally salient stimuli to draw attention. The time course of emotional effects on attention will be examined in much finer detail than previous work has allowed. The overall results should provide a broader and more realistic picture of how visual attention is controlled, and help identify processes that affect human performance in a variety of real-world tasks. Better understanding of these phenomena should prove useful in design of systems requiring rapid and accurate responses from people (e.g., airplane cockpits) and may have implications for eyewitness memory as well.

DESCRIPTION: The effects of retrieval practice on learning and memory have important implications, yet they have been investigated relatively little, and scarcely at all outside the narrow domain of verbal learning. Some data suggest that retrieval practice (without forced responding and with delayed correction) enhances learning more than the same amount of time devoted to additional study (Carrier and Pashler, 1992). If this principle has broad generality, it could be exploited to optimize training in many fields including medicine (e.g., mastering disease signs and symptoms, radiological diagnosis, etc.). New experiments and computational modeling are proposed to determine under what conditions retrieval practice increases learning, to assess the generality of testing effects beyond verbal learning, and to explore implications for underlying neural-computational mechanisms of learning. Testing inevitably leads to errors, but the consequences of errors have been little investigated. Edwin Guthrie argued that errors "stamp in" undesirable learning even if the person immediately realizes s/he made a mistake. Experiments will examine this intuitively appealing but untested notion, contrasting it with the alternative suggestion that errors may promote learning. A final set of experiments explores "fading" in perceptual discrimination training, starting with exaggerated differences and gradually narrowing the gap, thereby minimizing errors. A few studies suggest that fading speeds acquisition, but its power and generality are unknown. Experimental and computational studies are outlined to determine the possible benefits of fading and explore their possible implications for neural-computational bases of learning.

The general goal of this research program is to analyze the relationship between different attentional limitations and attentional mechanisms. We will use fine-grained chronometric analyses and converging experimental designs to address four theoretical issues. The first is the relationship between the "central bottleneck" that often arises in dual-task performance and the idea of a "central executive" believed responsible for the control and coordination of task sets. A tight linkage between these concepts is sometimes assumed but with little empirical basis. We hypothesize that they may be dissociable, and will test this by asking whether controlled shifts in task set can occur in parallel with central processing in concurrent tasks. The second issue to be explored concerns the possibility of parallel memory retrievals. The literature offers seemingly contradictory suggestions regarding whether multiple memory retrievals can operate at the same time. We will test a number of hypotheses that might reconcile these apparent conflicts. Third, we will explore the relationship between affectively charged stimuli and central attentional limitations, asking, e.g., whether such inputs elicit expressive and autonomic reactions independent of the central bottleneck and looking at the effect of emotional stimuli on scheduling of central processes. Finally, we will explore a hypothesis termed consonance-driven orienting that describes a general principle that might govern the interaction of central cognitive processes and perceptual attention. Recent work from our laboratory provides tentative support for this hypothesis; new experiments, both visual and auditory, are proposed to further test and refine it.

When someone 'crams' for a test, they often have the experience of forgetting the material soon after the test is over. In support of this experience, memory researchers have known for over 100 years that the long-term retention of study material can be greatly enhanced when study sessions are spaced out over time, rather than concentrated into one grueling cram session. Despite the robustness and importance of this spacing effect, there is currently no widely accepted explanation for it in terms of memory theories, and the breadth of the effect in terms of different kinds of study material has not been thoroughly investigated.

With support of the National Science Foundation, Dr. Pashler and Dr. Mozer are conducting a series of memory experiments to examine relations between inter-study intervals and retention intervals in the context of a variety of declarative memory tasks (i.e., fact learning) as well as procedural memory tasks (i.e., skill learning). The experiments are motivated by three distinct classes of memory theories that make different predictions regarding the performance outcomes. The investigators are implementing computational models of the theories to derive their quantitative predictions, and to explore their underlying theoretical principles. The knowledge to be gained from this research project may be used to improve learning and education in many academic domains, including science and engineering.

While traditional textbooks are designed to transmit information from the printed page to the learner, contemporary digital textbooks offer the opportunity to study learners as they interpret and process information being read. With a better understanding of a learner's state of mind, textbooks can make personalized recommendations for further study and review. How can the learner's state of mind be determined? Open a used printed textbook and the answer is clear: students feel compelled to engage with their texts by annotating key passages with highlights, tags, questions, and notes. Despite students' spontaneous desire to annotate as they read, this form of interaction has reaped few educational benefits in the past. At best, highlighted passages are re-read to study for exams, a strategy not nearly as effective as other strategies such as self-quizzing. This project will develop a new methodology that: assesses student knowledge level automatically based on annotations, transforms highlighted passages into appropriate study questions, and provides each student with well-timed, personalized review. Because the project is based on free, peer-reviewed, openly licensed materials from OpenStax that have been widely adopted at a range of institutions, particularly community colleges, the technology will reach beyond elite institutions to provide a broad spectrum of underserved students with access to a potentially powerful learning tool.

This project adopts a big-data approach that involves collecting annotations from a population of learners to draw inferences about individual learners. The project will determine how to exploit these data to model cognitive state, enabling the team to infer students' depth of understanding of facts and concepts, predict subsequent test performance, and perform interventions that improve learning outcomes. A tool will be developed that administers appropriately timed quizzes on material related to a student's highlights. A collaborative-filtering methodology will be employed that leverages population data to suggest specific passages for an individual to review. The proposed tool will reformulate selected passages into review questions that encourage the active reconstruction and elaboration of knowledge. The design and implementation of the tool will be informed by both randomized controlled studies within the innovative OpenStax textbook platform and coordinated laboratory studies. These studies will address basic scientific questions pertaining to why students annotate, how to improve their annotation skills, and techniques to optimize the use of annotations for guiding active review.