scholarly journals Provably-Robust Runtime Monitoring of Neuron Activation Patterns

Author(s):  
Chih-Hong Cheng
2021 ◽  
Vol 224 (9) ◽  
Author(s):  
Marine Banse ◽  
Boris P. Chagnaud ◽  
Alessia Huby ◽  
Eric Parmentier ◽  
Loïc Kéver

ABSTRACT In piranhas, sounds are produced through the vibration of the swim bladder wall caused by the contraction of bilateral sonic muscles. Because they are solely innervated by spinal nerves, these muscles likely evolved from the locomotor hypaxial musculature. The transition from a neuromuscular system initially shaped for slow movements (locomotion) to a system that requires a high contraction rate (sound production) was accompanied with major peripheral structural modifications, yet the associated neural adjustments remain to this date unclear. To close this gap, we investigated the activity of both the locomotor and the sonic musculature using electromyography. The comparison between the activation patterns of both systems highlighted modifications of the neural motor pathway: (1) a transition from a bilateral alternating pattern to a synchronous activation pattern, (2) a switch from a slow- to a high-frequency regime, and (3) an increase in the synchrony of motor neuron activation. Furthermore, our results demonstrate that sound features correspond to the activity of the sonic muscles, as both the variation patterns of periods and amplitudes of sounds highly correspond to those seen in the sonic muscle electromyograms (EMGsonic). Assuming that the premotor network for sound production in piranhas is of spinal origin, our results show that the neural circuit associated with spinal motor neurons transitioned from the slow alternating pattern originally used for locomotion to a much faster simultaneous activation pattern to generate vocal signals.


Author(s):  
Run Wang ◽  
Felix Juefei-Xu ◽  
Lei Ma ◽  
Xiaofei Xie ◽  
Yihao Huang ◽  
...  

In recent years, generative adversarial networks (GANs) and its variants have achieved unprecedented success in image synthesis. They are widely adopted in synthesizing facial images which brings potential security concerns to humans as the fakes spread and fuel the misinformation. However, robust detectors of these AI-synthesized fake faces are still in their infancy and are not ready to fully tackle this emerging challenge. In this work, we propose a novel approach, named FakeSpotter, based on monitoring neuron behaviors to spot AI-synthesized fake faces. The studies on neuron coverage and interactions have successfully shown that they can be served as testing criteria for deep learning systems, especially under the settings of being exposed to adversarial attacks. Here, we conjecture that monitoring neuron behavior can also serve as an asset in detecting fake faces since layer-by-layer neuron activation patterns may capture more subtle features that are important for the fake detector. Experimental results on detecting four types of fake faces synthesized with the state-of-the-art GANs and evading four perturbation attacks show the effectiveness and robustness of our approach.


2016 ◽  
Vol 224 (2) ◽  
pp. 62-70 ◽  
Author(s):  
Thomas Straube

Abstract. Psychotherapy is an effective treatment for most mental disorders, including anxiety disorders. Successful psychotherapy implies new learning experiences and therefore neural alterations. With the increasing availability of functional neuroimaging methods, it has become possible to investigate psychotherapeutically induced neuronal plasticity across the whole brain in controlled studies. However, the detectable effects strongly depend on neuroscientific methods, experimental paradigms, analytical strategies, and sample characteristics. This article summarizes the state of the art, discusses current theoretical and methodological issues, and suggests future directions of the research on the neurobiology of psychotherapy in anxiety disorders.


2005 ◽  
Vol 32 (S 4) ◽  
Author(s):  
A.R Luft ◽  
L Forrester ◽  
F Villagra ◽  
R Macko ◽  
D.F Hanley

Sign in / Sign up

Export Citation Format

Share Document