SignificanceThe earliest way humans can learn what their body is and where the outside world begins is through the tactile sense, especially through touch between parent and baby. In this study, we demonstrated differential processing of touch from self and others at cortical and spinal levels. Our results support top-down modulation of dorsal horn somatosensory processing, as recently shown in animal studies. We provide evidence that the individual self-concept relates to differential self- vs. other-processing in the tactile domain. Self- vs. other-distinction is necessary for successful social interaction with others and for establishing a coherent self. Our results suggest an association between impaired somatosensory processing and a dysfunctional self-concept, as seen in many psychiatric disorders.

Grasping and manipulating an object requires us to perceive its material compliance. Compliance is thought to be encoded by relationships of force, displacement, and contact area at the finger pad. Prior work suggests that objects must be sufficiently deformed to become discriminable, but the utility of time-dependent cues has not been fully explored. The studies herein find that the availability of force-rate cues improve compliance discriminability so as to require less deformation of stimulus and finger pad. In particular, we tested the impact of controlling force-rate and displacement-rate cues in passive touch psychophysical experiments. An ink-based method to mark the finger pad was used to measure contact area per stimulus, simultaneously with displacement and force. Compliances spanned a range harder and softer than the finger pad. The results indicated harder compliances were discriminable at lower peak forces when the stimulus control mode was displacement-rate (0.5 N) compared to force-rate (1.3 N). That is, when displacement-rate was controlled to be equal between the two compliances, the resultant force-rate psychophysical cues could be more readily discriminated. In extending prior studies, while some magnitude of finger pad deformation may be sufficient for discriminability, temporal cues tied to force afford more efficient judgments.

No abstract

No abstract

Applications of machine learning are subject to three major components that contribute to the final performance metrics. Within the category of neural networks, and deep learning specifically, the first two are the architecture for the model being trained and the training approach used. This work focuses on the third component, the data used during training. The primary questions that arise are “what is in the data” and “what within the data matters?” looking into the radio frequency machine learning (RFML) field of automatic modulation classification (AMC) as an example of a tool used for situational awareness, the use of synthetic, captured, and augmented data are examined and compared to provide insights about the quantity and quality of the available data necessary to achieve desired performance levels. Three questions are discussed within this work: (1) how useful a synthetically trained system is expected to be when deployed without considering the environment within the synthesis, (2) how can augmentation be leveraged within the RFML domain, and, lastly, (3) what impact knowledge of degradations to the signal caused by the transmission channel contributes to the performance of a system. In general, the examined data types each make useful contributions to a final application, but captured data germane to the intended use case will always provide more significant information and enable the greatest performance. Despite the benefit of captured data, the difficulties and costs that arise from live collection often make the quantity of data needed to achieve peak performance impractical. This paper helps quantify the balance between real and synthetic data, offering concrete examples where training data is parametrically varied in size and source.