Visual search and childhood vision impairment: A GAMLSS-oriented multiverse analysis approach

The aim of this report was to analyze reaction times and accuracy in children with a vision impairment performing a feature-based visual search task using a multiverse statistical approach. The search task consisted of set sizes 4, 16, and 24, consisting of distractors (circle) and a target (ellipse) that were presented randomly to school-aged individuals with or without a vision impairment. Interactions and main effects of key variables relating to reaction times and accuracy were analyzed via a novel statistical method blending GAMLSS (generalized additive models for location, scale, and shape) and distributional regression trees. Reaction times for the target-present and target-absent conditions were significantly slower in the vision impairment group with increasing set sizes (p < .001). Female participants were significantly slower than were males for set sizes 16 and 24 in the target-absent condition (p < .001), with male participants being significantly slower than females in the target-present condition (p < .001). Accuracy was only significantly worse (p = .03) for participants less than 14 years of age for the target-absent condition with set sizes 16 and 24. There was a positive association between binocular visual acuity and search time (p < .001). The application of GAMLSS with distributional regression trees to the analysis of visual search data may provide further insights into underlying factors affecting search performance in case-control studies where psychological or physical differences may influence visual search outcomes.

Oddball onset timing: Little evidence of early gating of oddball stimuli from tapping, reacting, and producing

Oddballs, rare or novel stimuli, appear to last longer than non-oddballs. This illusion is often attributed to the perceived time that an oddball occupies being longer than that of a non-oddball. However, it is also possible that oddball stimuli are perceived to onset earlier than non-oddballs; they are “gated” earlier in time and thus the perceived duration of those stimuli are longer. In the current article, we directly investigate this proposal by asking participants to react to, produce durations initiated with, and tap along to either oddball or standard stimuli. Tapping provided some support for earlier perceived onset of an oddball in the visual modality. However, both reaction time and duration production experiments provided evidence against an oddball being gated earlier than a standard stimulus. Contrarily, these experiments showed an oddball resulted in longer reaction times and productions, respectively. Taken together, these three experiments indicate it is unlikely that the expansion of time attributed to oddball presentation is purely due to the earlier gating of oddball stimuli. In fact, the first two experiments provide some evidence that the effect of an oddball must compensate for the later gating of these stimuli.

A beginning quantitative taxonomy of cognitive activation systems and application to continuous flow processes

Much progress has been made in the investigation of perceptual, cognitive, and action mechanisms under the assumption that when one subprocess precedes another, the first one starts and finishes before the other begins. We call such processes "Dondersian" after the Dutch physiologist who first formulated this concept. Serial systems obey this precept (e.g., Townsend, 1974). However, most dynamic systems in nature do not: instead, each subprocess communicates its state to its immediate successors continuously. Although the mathematics for physical systems has received extensive treatment over the last three centuries, applications to human cognition have been exiguous. Therefore, the pioneering papers by Charles Eriksen and colleagues on continuous flow dynamics (e.g., Eriksen & Schulz, Perception & Psychophysics, 25, 249-263, 1979; Coles et al.,, Journal of Experimental Psychology: Human Perception and Performance, 11(5), 529, 1985) must be viewed as truly revolutionary. Surprisingly, there has been almost no advancement on this front since. With the goal of bringing this theme back into the scientific consciousness and extending and deepening our understanding of such systems, we develop a taxonomy that emphasizes the fundamental characteristics of continuous flow dynamics. Subsequently, we complexify the treated systems in such a way as to illustrate the popular cascade model (Ashby, Psychological Review, 89, 599-607, 1982; McClelland, Psychological Review, 86, 287-330, 1979) and use it to simulate the classic findings of Eriksen and colleagues (Eriksen & Hoffman, Perception & Psychophysics, 12(2), 201-204, 1972).

A Limiting Channel Capacity of Visual Perception: Spreading Attention Divides the Rates of Perceptual Processes

This study investigated effects of divided attention on the temporal processes of perception. During continuous watch periods, observers responded to sudden changes in the color or direction of any one of a set of moving objects. The set size of moving objects was a primary variable. A simple detection task required responses to any display change, and a selective task required responses to a subset of the changes. Detection rates at successive points in time were measured by response time (RT) hazard functions.The principal finding was that increasing the set size divided the detection rates-and these divisive effects were essentially constant over time and over the time-varying influence of the target signals and response tasks. The set size, visual target signal, and response task exerted mutually invariant influence on detection rates at given times, indicating independent joint contributions of parallel component processes. The lawful structure of these effects was measured by RT hazard functions but not by RTs as such. The results generalized the time-invariant divisive effects of set size on visual process rates found by Lappin, Morse, & Seiffert (Attention, Perception, & Psychophysics, 78, 2469-2493, 2016). These findings suggest that the rate of visual perception has a limiting channel capacity.

Forget binning and get SMART: Getting more out of the time-course of response data

Many experiments aim to investigate the time-course of cognitive processes while measuring a single response per trial. A common first step in the analysis of such data is to divide them into a limited number of bins. As we demonstrate here, the way one chooses these bins can considerably influence the resulting time-course. As a solution to this problem, we here present the smoothing method for analysis of response time-course (SMART)-a complete package for reconstructing the time-course from one-sample-per-trial data and performing statistical analysis. After smoothing the data, the SMART weights the data based on the effective number of data points per participant. A cluster-based permutation test then determines at which moments the responses differ from a baseline or between two conditions. We show here that, in contrast to contemporary binning methods, the chosen temporal resolution has a negligible effect on the SMART reconstructed time-course. To facilitate its use, the SMART method, accompanied by a tutorial, is available as an open-source package.

How many trials are required for parameter estimation in diffusion modeling? A comparison of different optimization criteria

Diffusion models (Ratcliff, 1978) make it possible to identify and separate different cognitive processes underlying responses in binary decision tasks (e.g., the speed of information accumulation vs. the degree of response conservatism). This becomes possible because of the high degree of information utilization involved. Not only mean response times or error rates are used for the parameter estimation, but also the response time distributions of both correct and error responses. In a series of simulation studies, the efficiency and robustness of parameter recovery were compared for models differing in complexity (i.e., in numbers of free parameters) and trial numbers (ranging from 24 to 5,000) using three different optimization criteria (maximum likelihood, Kolmogorov-Smirnov, and chi-square) that are all implemented in the latest version of fast-dm (Voss, Voss, & Lerche, 2015). The results revealed that maximum likelihood is superior for uncontaminated data, but in the presence of fast contaminants, Kolmogorov-Smirnov outperforms the other two methods. For most conditions, chi-square-based parameter estimations lead to less precise results than the other optimization criteria. The performance of the fast-dm methods was compared to the EZ approach (Wagenmakers, van der Maas, & Grasman, 2007) and to a Bayesian implementation (Wiecki, Sofer, & Frank, 2013). Recommendations for trial numbers are derived from the results for models of different complexities. Interestingly, under certain conditions even small numbers of trials (N < 100) are sufficient for robust parameter estimation.

Experimental validation of the diffusion model based on a slow response time paradigm

The diffusion model (Ratcliff, Psychol Rev 85(2):59-108, 1978) is a stochastic model that is applied to response time (RT) data from binary decision tasks. The model is often used to disentangle different cognitive processes. The validity of the diffusion model parameters has, however, rarely been examined. Only few experimental paradigms have been analyzed with those being restricted to fast response time paradigms. This is attributable to a recommendation stated repeatedly in the diffusion model literature to restrict applications to fast RT paradigms (more specifically, to tasks with mean RTs below 1.5 s per trial). We conducted experimental validation studies in which we challenged the necessity of this restriction. We used a binary task that features RTs of several seconds per trial and experimentally examined the convergent and discriminant validity of the four main diffusion model parameters. More precisely, in three experiments, we selectively manipulated these parameters, using a difficulty manipulation (drift rate), speed-accuracy instructions (threshold separation), a more complex motoric task (non-decision time), and an asymmetric payoff matrix (starting point). The results were similar to the findings from experimental validation studies based on fast RT paradigms. Thus, our experiments support the validity of the parameters of the diffusion model and speak in favor of an extension of the model to paradigms based on slower RTs.

The role of unique color changes and singletons in attention capture

Previous studies have shown that a sudden color change is typically less salient in capturing attention than the onset of a new object. Von Mühlenen, Rempel, and Enns (Psychological Science 16: 979-986, 2005) showed that a color change can capture attention as effectively as the onset of a new object given that it occurs during a period of temporal calm, where no other display changes happen. The current study presents a series of experiments that further investigate the conditions under which a change in color captures attention, by disentangling the change signal from the onset of a singleton. The results show that the item changing color receives attentional priority irrespective of whether this change goes along with the appearance of a singleton or not.

Simple reaction time for broadband sounds compared to pure tones

Although many studies have explored the relation between reaction time (RT) and loudness, including effects of intensity, frequency, and binaural summation, comparable work on spectral summation is rare. However, most real-world sounds are not pure tones and typically have bandwidths covering several critical bands. Since comparing to a 1-kHz pure tone, the reference tone, is important for loudness measurement and standardization, the present work focuses on comparing RTs for broadband noise to those for 1-kHz pure tones in three experiments using different spectral and binaural configurations. The results of Experiments 1 and 2 yield good quantitative agreement with spectral loudness summation models for moderate and high sound pressure levels, measured using both pink noise covering almost the entire hearing range and bandpass-filtered pink noise with different center frequencies. However, at lower levels, the RT measurements yield an interaction of level and bandwidth, which is not in line with loudness scaling data. In Experiment 3, which investigated the binaural summation of broadband sounds, the binaural gain for white noise was determined to be 9 dB, which is somewhat larger than what had been found in previous RT measurements using 1-kHz pure tones.

Big and small numbers: Empirical support for a single, flexible mechanism for numerosity perception

The existence of perceptually distinct numerosity ranges has been proposed for small (i.e., subitizing range) and larger numbers based on differences in precision, Weber fractions, and reaction times. This raises the question of whether such dissociations reflect distinct mechanisms operating across the two numerosity ranges. In the present work, we explore the predictions of a single-layer recurrent on-center, off-surround network model of attentional priority that has been applied to object individuation and enumeration. Activity from the network can be used to model various phenomena in the domain of visual number perception based on a single parameter: the strength of inhibition between nodes. Specifically, higher inhibition allows for precise representation of small numerosities, while low inhibition is preferred for high numerosities. The model makes novel predictions, including that enumeration of small numerosities following large numerosities should result in longer reaction times than when a small numerosity trial following small numerosities. Moreover, the model predicts underestimation of number when a display containing a large number of items follows a trial with small numerosities. We behaviorally confirmed these predictions in a series of experiments. This pattern of results is consistent with a single, flexible object individuation system, which can be modeled successfully by dynamic on-center, off-surround network model of the attentional priority (saliency) map.

Retest reliability of the parameters of the Ratcliff diffusion model

In the recent years, there is a growing interest to use the Ratcliff Diffusion Model (1978) for diagnostic purposes as the parameters of the model capture interindividual differences in specific cognitive processes. The parameters are estimated using reaction time data from binary classification tasks. For a potential diagnostic application of parameter values sufficient reliability is a necessary precondition. In two studies, each with two sessions separated by 1 week, the retest reliability of the diffusion model parameters was assessed. In Study 1, 105 participants completed a lexical decision task and a recognition memory task. In Study 2, 128 participants worked on an associative priming task. Results show that the reliability of the main parameters of the Ratcliff Diffusion Model (in particular of the speed of information accumulation and the threshold separation with rs > 0.70 for all three tasks) is satisfying. Besides, we analyzed the influence of the number of trials on the retest reliability using different estimation methods (Kolmogorov-Smirnov, Maximum Likelihood, Chi-square and EZ) and both empirical and simulated datasets.

Simple reaction time to the onset of time-varying sounds

Although auditory simple reaction time (RT) is usually defined as the time elapsing between the onset of a stimulus and a recorded reaction, a sound cannot be specified by a single point in time. Therefore, the present work investigates how the period of time immediately after onset affects RT. By varying the stimulus duration between 10 and 500 msec, this critical duration was determined to fall between 32 and 40 milliseconds for a 1-kHz pure tone at 70 dB SPL. In a second experiment, the role of the buildup was further investigated by varying the rise time and its shape. The increment in RT for extending the rise time by a factor of ten was about 7 to 8 msec. There was no statistically significant difference in RT between a Gaussian and linear rise shape. A third experiment varied the modulation frequency and point of onset of amplitude-modulated tones, producing onsets at different initial levels with differently rapid increase or decrease immediately afterwards. The results of all three experiments results were explained very well by a straightforward extension of the parallel grains model (Miller and Ulrich Cogn. Psychol. 46, 101-151, 2003), a probabilistic race model employing many parallel channels. The extension of the model to time-varying sounds made the activation of such a grain depend on intensity as a function of time rather than a constant level. A second approach by mechanisms known from loudness produced less accurate predictions.