The story of information processing is a story of great success. Todays' microprocessors are devices of unprecedented complexity and MOSFET transistors are considered as the most widely produced artifact in the history of mankind. The current miniaturization of electronic circuits is pushed almost to the physical limit and begins to suffer from various parasitic effects. These facts stimulate intense research on neuromimetic devices. This feature article is devoted to various in materio implementation of neuromimetic processes, including neuronal dynamics, synaptic plasticity, and higher-level signal and information processing, along with more sophisticated implementations, including signal processing, speech recognition and data security. Due to vast number of papers in the field, only a subjective selection of topics is presented in this review.
The story of information processing is a story of great success. Todays' microprocessors are devices of unprecedented complexity and MOSFET transistors are considered as the most widely produced component in the history of mankind, with a total number of devices approaching 13 sextillions (1.3×10 22 ) of fabricated devices. 1 The contemporary microprocessor contains approximately 3.95×10 9 transistors. 2 In a human, there are an estimated 10-20 billion neurons in the cerebral cortex and 55-70 billion neurons in the cerebellum. 3 A human brain is, therefore, the most complex information processing structure, as each of the 6×10 10 neurons may form up to 10 4 synaptic connections with other neurons. Brain's ability to learn and adapt is a consequence if its dynamically changing topology of synaptic connections, plasticity of individual connections, high redundancy and multilevel dynamics at various geometrical and temporal scales.
These facts stimulate intense research on neuromimetic devices. Their performance, at the present stage of development, cannot be compared with natural systems, but also provide stimulation for other fields of investigation, including chemistry, physics, electronics, and computer sciences.
This feature article is devoted to various in materio implementation of neuromimetic processes, including neuronal dynamics, synaptic plasticity, and higher-level signal and information processing. From the plethora of various in-materio implementation of information processing, involving inorganic and organic materials, polymers, various molecular species, as well as biopolymers and even living organisms 4 we have chosen a handful of wet photochromic systems and semiconducting materials. This selection is by no means exhaustive, but sufficient to illustrate the main research directions as well as current trends in in-materio neuromimetic computing.
Human intelligence emerges from the complex structural and dynamical properties of our nervous system. The primary cellular elements of our nervous systems are neurons. The ultimate computational power of our nervous system relies on the dynamical properties of neurons and their networks. Every neuron is a nonlinear dynamic system, 5 and according to some theoretical analysis can be regarded as a biological memristive element. [6][7][8] Some neurons operate in the oscillatory regime. They are called pacemaker neurons and fire action potentials periodically. Pacemaker neurons generate rhythmic activities in neural networks involved in the neocortex, basal ganglia, thalamus, locus coeruleus, hypothalamus, ventral tegmentum area, hippocampus, and amygdala. 9 These structures are associated with sleep, wakefulness, arousal, motivation, addiction, memory consolidation, cognition, and fear.
Excitable neurons are another type of neurons present in the nervous system. Excitability can be twofold -“tonic” or “phasic”. When neurons react to a constant excitatory signal by firing a sequence of spikes, they are classified as “tonic”. On the other hand, excitability is “phasic”, when neurons react in an analog manner and shoot only once, when receiving a sharp excitatory signal. Excitatory “tonic” neurons are present e.g. in the cortex whereas “phasic” excitable neurons act e.g. in the auditory brainstem (involved in precise timing computations) and in the spinal cord. 10,11 Finally, there are chaotic neurons. Chaotic neurons are quite common in the nervous system because the intrinsic dynamic instability facilitates the extraordinary ability of neural networks to adapt. 12,13 It is possible to mimic the dynamics of neurons by selecting specific chemical systems and maintaining them out-of-equilibrium. 14 One of the most widespread examples is the Belousov-Zhabotinsky reaction. Other popular models are memristive elements and circuits, [6][7][8] as well as some (photo)electrochemical systems, especially those with self-excitable oscillations. 15
The Belousov-Zhabotinsky (BZ) reaction is a catalyzed oxidative bromination of malonic acid in aqueous acidic solution (1): 2BrO3 - (aq) + 3CH2(COOH)2(aq) + 2H + (aq) → 2BrCH(COOH)2(aq) + 3CO2(g) + 4H2O(l) (1) Various metal ions or metal-complexes, such as either cerium ions or ferroin (i.e. tris-(1,10-phenanthroline)-iron(II)) or tris(2,2’-bipyridyl)dichloro-ruthenium(II) (Ru(bpy)3 2+ ) can serve as catalyst. The mechanism of the BZ is quite complicated because it consists of many elementary steps. Briefly, when the concentration of the intermediate bromide (𝐵𝐵𝑟𝑟 -) is higher than its critical value, the reaction proceeds by a set of elementary steps wherein the catalyst maintains the reduced state, and the solution is red-colored in the presence of ferroin. During these elementary steps, bromide is consumed. As soon as the concentration of 𝐵𝐵𝑟𝑟 -is lower than its critical value (that corresponds to 5x10 -6 [𝐵𝐵𝑟𝑟𝑂𝑂 3 -]), the reaction proceeds through another set of elementary steps, where mono-electronic transformati
This content is AI-processed based on open access ArXiv data.
