Three types of point estimates of parameters, means, medians, and maximum a posteriori (MAP) estimates, were compared with respect to bias and root mean square error (RMSE). The overall results showed that the MAP estimates were the best among the three estimators for asymmetric posterior distributions of proportion parameters of binomial models, standard deviation parameters of univariate normal models, and correlation coefficient parameters of bivariate normal models. Although the comparisons were made for simple models, the results suggested that, in general, MAP estimates are appropriate for true values because MAP estimates are included in any highest-density intervals and because MAP estimates with flat prior distributions coincide with maximum likelihood estimates, which asymptotically converge to true values when sample sizes become large. A simple Python script to calculate a MAP estimate from Markov chain Monte Carlo (MCMC) sampling was presented.

Literature suggests observers can extract an average of facial expressions or attractiveness of multiple faces presented simultaneously. This phenomenon is called “ensemble perception.” However, it is unclear whether observers identify the averaged appearance of faces in ensemble perception. To address this, we conducted two experiments where participants studied a set of two different faces as targets and were then asked to judge whether the test face was an average of the two targets. We prepared three types of test faces: an averaged face using the two targets (T2), an averaged face using the target and non-target faces (TN), and an averaged face using two non-target faces (N2). In both experiments, participants correctly judged T2 as targets as opposed to N2. However, participants could discriminate T2 from TN only in Experiment 2, where correct or incorrect feedback at each trial was provided. These results suggest that observers can identify an averaged face from two targets but discriminate the averaged one from others only when facilitated by feedback.

The world looks vividly three-dimensional when we use the two eyes; every object has thickness, occupies a volume, and is separated in depth from others. This perception largely relies on processing of binocular disparity, a small horizontal shift between the projections of each visual feature onto the left and right retinae. Binocular disparity is detected in the primary visual cortex (V1) by a process similar to calculation of cross-correlation between the left and right retinal images. The neural signals from V1 are then processed along both dorsal and ventral pathways. Dorsal pathway areas MT and MST represent absolute disparity as in V1, and mediate coarse stereopsis and reflexive vergence eye movement. Ventral pathway areas V4 and IT compute relative disparity between features and are involved in fine stereopsis. V4 and IT convert the correlation-based signals into disparity representations of binocularly matched features, solving the stereo correspondence problem. MT neurons shows responses intermediate between the correlation-based and match-based representations. The two pathways thus contribute to stereo perception in a complementary and parallel manner.

Since Deep Convolutional Neural networks (DCNs) have achieved excellent general object recognition performance, researchers in the fields of systems neuroscience and visual psychophysics have attempted to understand visual processing in the brain via DCNs. In this talk, I will review the progress of DCNs in computer vision, and introduce the authors’ visual neuroscience studies using DCNs. Then, I will discuss the prospects for multi-omics analysis of psychonomic science data based on recent research trends in deep learning and multimodal processing.

Visual illusions are phenomena in which the actual physical parameters of an object are out of alignment with visual perception, such as when an object that is not moving appears to move, or when the color, brightness, or size of an object appears to be different from its reality. It is also the perception of such discrepancies. Illusions have attracted many people for a long time because they are surprising to the viewer and have a strong entertainment value. Illusions are thought to express the characteristics of visual information processing in the eye and brain, and are an important tool in the study of vision, and the field of research has been developing remarkably. In recent years, research methods have actively incorporated not only psychophysical methods but also modern brain science methodologies such as fMRI, and more recently, research utilizing artificial intelligence has also progressed. This paper focuses on the study of visual illusions using deep neural networks (DNN),which is being conducted in the authors’ laboratory and is becoming popular worldwide.

To understand the experience of seeing is difficult and is being pursued day and night, especially in psychology and neuroscience. However, there are hard areas to research, such as estimating the structure of receptive fields (RF) in the higher-order visual cortex with humans and animals as research targets. Deep neural networks (DNNs) are being reported that have similar properties to visual neurons and the possibility of using DNNs as alternative research targets to the biological brain has emerged. Therefore, in this paper, I discuss whether DNNs can be our research subject. In this research, I applied the reverse correlation method, which has revealed RF of visual neurons, to DNNs to estimate RF of units in well-trained VGG-16. As a result, the properties of the RF of VGG-16 units were similar to visual cortex neurons. The result suggested that DNNs may be a good alternative model for our research, but also suggested limitations of the method. To solve remaining problems, psychology research that develop the better methods and deep learning research that provides better alternative models must go hand in hand.

When conducting experimental psychology, psychological phenomena are often described in mathematical terms and logic. How practical is the mathematical approach in experimental psychology? This paper aims to explain plain linguistic and mathematical expressions, how psychological phenomena are expressed mathematically in experimental psychology and how mathematical models are constructed, using examples of psychological phenomena related to sensation, perception, and cognition. First, it outlined the perspective, thinking, and utilization of mathematical models in experimental psychology, using some psychological phenomena. Second, it was introduced that several previous studies dealt with spatial properties of perceptual and cognitive phenomena, including the author’s studies. It is hoped that this paper will increase familiarity with and opportunities to utilize mathematical models in experimental psychology.

Our study utilized mathematical modeling to explore the mechanism of the perceptual deterioration for self-produced intervals. A psychophysical experiment showed that perceptual precision for self-produced intervals was lower than that for passively presented intervals. We hypothesized that common noise between perceptual and motor timing was a critical factor in this perceptual deterioration. A Mathematical model based on this common hypothesis showed that a higher level of the common noise leads to lower perceptual precision. Thus, we believe that mathematical modeling is a useful tool that can strengthen conceptual hypotheses and deepen our understanding of psychological phenomena by combining them with experiments.

Optimal foraging theory is a branch of behavioral ecology. This specific set of models tackles questions about how an animal should perform a behavior in a prey-searching situation. In this article, I argue that cognitive psychologists can also benefit from using visual foraging tasks to study human behaviors when searching for visual information. Specifically, by modeling information about the search environment as a gain function and incorporating it into the model, the model of optimal patch use can be used to predict human search behavior. The use of such modeling is expected to lead to the development of visual search research by more clearly defining the assumptions and theory of the behavior to be examined. I also note that the activities of young researchers in cognitive modeling have been particularly remarkable in recent years, and further development is expected in the future.

Virtual reality (VR) is a new methodology for psychological studies. Especially in modern VR head-mounted displays (HMDs), researchers can control more complex and dynamic stimulus presentation than in standard laboratory experiments, enhancing the ecological validity of research. While the studies by VR HMDs have increased, little is known about what are technical points to establish more stable VR environments and whether current VR HMDs have millisecond accuracy and precision for stimulus presentation compared to traditional laboratory environments, since most standard methods in psychological studies are not optimized for VR environments. Thus, the present article provides key features for setup of more reliable VR studies, introducing recent works that have revealed the time/timing accuracy and precision of visual and auditory stimulus presentation in modern VR HMDs.