The Demon Hidden Behind Life's Ultra-Energy-Efficient Information Processing -- Demonstrated by Biological Molecular Motors

TL;DR

The paper addresses the mismatch between the energy demands of modern digital AI and the ultra-efficient information processing observed in biology. It experimentally demonstrates a Maxwell's demon–like information–energy conversion in a biological molecular motor by directly linking positional information to mechanical work, using high-resolution single-molecule tracking of a DNA-origami–reconstituted myosin-on-actin system (0.7 nm spatial, 40 μs temporal resolution). The authors quantify an information–energy mapping with $W = k_B T ln 2 × I$, show that rare Brownian fluctuations with probability ~1/3000 can yield approximately 40% efficiency per ATP turnover within ~300 ms cycles, and estimate $I ≈ 11$ bits per cycle, bridging microscopic fluctuations to macroscopic work. They argue this demon-like processing is a general principle across biological molecular machines and outline design rules for noise-driven, energy-efficient AI inspired by biology, potentially informing nanoscale actuators and brain-like computation.

Abstract

The remarkable progress of artificial intelligence (AI) has revealed the enormous energy demands of modern digital architectures, raising deep concerns about sustainability. In stark contrast, the human brain operates efficiently on only ~20 watts, and individual cells process gigabit-scale genetic information using energy on the order of trillionths of a watt. Under the same energy budget, a general-purpose digital processor can perform only a few simple operations per second. This striking disparity suggests that biological systems follow algorithms fundamentally distinct from conventional computation. The framework of information thermodynamics-especially Maxwell's demon and the Szilard engine-offers a theoretical clue, setting the lower bound of energy required for information processing. However, digital processors exceed this limit by about six orders of magnitude. Recent single-molecule studies have revealed that biological molecular motors convert Brownian motion into mechanical work, realizing a "demon-like" operational principle. These findings suggest that living systems have already implemented an ultra-efficient information-energy conversion mechanism that transcends digital computation. Here, we experimentally establish a quantitative correspondence between positional information (bits) and mechanical work, demonstrating that molecular machines selectively exploit rare but functional fluctuations arising from Brownian motion to achieve ATP-level energy efficiency. This integration of information, energy, and timescale indicates that life realizes a Maxwell's demon-like mechanism for energy-efficient information processing.

Figures (5)

• Figure 1: Comparison of energy consumption between artificial neural networks and biological brains.
• Figure 2: Szilard engine and Maxwell's demon. The demon observes a single molecule undergoing Brownian motion in a space connected to a heat bath (1). When the molecule is detected to have randomly moved to the right side of the space (2), the demon acquires and stores this information (1 bit) and then inserts a partition (piston) at the center (3). The partition is pushed leftward by collisions with the molecule, and the molecule expands isothermally into the available space. Through this isothermal expansion, the system performs work of $k_{\mathrm{B}}T\ln 2$ on the surroundings (4). Subsequently, the demon erases its memory by consuming $k_{\mathrm{B}}T\ln 2$ of energy, thereby returning the system to its initial state (5).
• Figure 3: Molecular motors in muscle.
• Figure 4: High-resolution single-molecule measurement of the molecular motor myosin. A myosin head (molecular motor) protruding via an elastic element (spring) from a DNA-origami-based myosin filament backbone is allowed to interact with an actin filament immobilized on a glass surface in the presence of ATP. The dynamics of the myosin head are observed by tracking the position of a gold nanoparticle selectively attached to it. The gold nanoparticle is illuminated by the evanescent field generated through total internal reflection of a green laser at the glass surface (TIRF illumination), and the resulting scattered light image is captured with a high-speed camera. Image analysis enables measurement of the centroid position of the gold nanoparticle—and thereby of the molecular motor—with a spatial resolution of $0.7~\mathrm{nm}$ and a temporal resolution of $40~\mu\mathrm{s}$Fujita2019.
• Figure 5: Molecular motor converting positional information into motion by exploiting Brownian fluctuations. The dynamics of a molecular motor measured by high-resolution single-molecule analysis (Fig. \ref{['fig:fig4']}) are shown. The central trace represents the actual trajectory of the observed molecular motor. The motor repeatedly binds to and dissociates from the actin filaments on a microsecond timescale, undergoing forward and backward Brownian motion. When it happens to fluctuate forward far enough to contact the next actin subunit, the strain sensor of the motor is activated and binding occurs (2). Upon release of ADP and Pi generated by ATP hydrolysis, the motor undergoes a structural transition and binds strongly and stably to the forward actin subunit (3). As it moves, the elastic element (spring) connecting the extended myosin head to the myosin filament generates force between the two filaments (4). When a new ATP molecule (energy carrier) binds to the motor, the strong actin attachment is released, and Brownian motion resumes (5). The circled numbers correspond to the processes illustrated in Fig. \ref{['fig:fig2']}. The third panel (right) shows part of a high-speed AFM movie directly capturing the structural change (tail bending) of the myosin head during step from (2) to (3) Fujita2019.