Source-linked AI summary
Possible existence of optical communication channels in the brain
Sourabh Kumar, Kristine Boone, Jack Tuszynski, Paul E. Barclay, Christoph Simon
TL;DR
The paper asks whether biophotons could supplement electro-chemical signaling, with targeted optical communication requiring waveguides. It numerically models light propagation in myelinated axons and finds that bend transmission is comparable in longer and shorter segments at the same change in curvature, while noting unresolved limits in the optical assumptions.
Problem
The paper asks whether biophotons could carry information between neurons and what structures could guide them through the brain.
Source: Discussion · p. 9Method
The authors numerically model light propagation through myelinated axons, including paranodal geometry and bend-related transmission.
Source: Discussion · pp. 9–10Results
For bends, 150 µm segments had transmission comparable to shorter segments at the same ∆κ, supporting change of curvature as the key loss factor.
Source: Discussion · p. 10Takeaways & Limitations
The authors conclude that light conduction in myelinated axons is possible with realistic imperfections and propose experiments to test the hypothesis.
Source: Discussion · p. 9The model assumes a uniform axonal refractive index even though observed values span 1.34–1.40.
Source: in the brain” · p. 20
Abstract
from arXiv · showhide
Given that many fundamental questions in neuroscience are still open, it seems pertinent to explore whether the brain might use other physical modalities than the ones that have been discovered so far. In particular it is well established that neurons can emit photons, which prompts the question whether these biophotons could serve as signals between neurons, in addition to the well-known electro-chemical signals. For such communication to be targeted, the photons would need to travel in waveguides. Here we show, based on detailed theoretical modeling, that myelinated axons could serve as photonic waveguides, taking into account realistic optical imperfections. We propose experiments, both \textit{in vivo} and \textit{in vitro}, to test our hypothesis. We discuss the implications of our results, including the question whether photons could mediate long-range quantum entanglement in the brain.
Results
Numerical modeling indicates that myelinated axons can guide light despite nodal structure and realistic optical irregularities, while transmission depends on wavelength and axon geometry.
Nodal, and paranodal regions
- At p-ratio ≈2.5, shorter wavelengths transmit better, while longer wavelengths fare better at larger p-ratios; caliber trends also reverse under some conditions.Thicker axons usually transmit better at p-ratio ≈2.5, but thinner axons can do better for short wavelengths and p-ratios around 5 or greater; transmission across multiple paranodes is approximately predicted by exponentiating single-region transmission.Source: Results · p. 4
Bends
- Shorter wavelengths are better confined and transmit more through bends; for thin axons, some measured 1 mm segments with ∆κ < 0.05 µm^-1 yield greater than 90% transmission.The inner part of a bent axon experiences greater curvature than the outer part, so thicker axons can suffer more loss at the same ∆κ.Source: Results · pp. 4–5
Varying cross-sectional area; Non-circular cross section
- Greater radius inhomogeneity and cross-sectional irregularity reduce transmission, while all axons have close to unit efficiency for less than 10% radius variation.Thinner axons tolerate greater percentage inhomogeneity; longer roughness correlation lengths improve transmission at the same variation s.d., and the cross-section results concern coupling loss when shape stays stable along the axon.Source: Results · pp. 5, 7
Other imperfections
- Close side-by-side axons can exchange light, so the modeled guidance generally requires about a wavelength of separation to prevent cross-talk.The simulations considered identical axons, for which cross-talk is stronger than between non-identical axons; its effects on longer lengths are not straightforward.Source: Results · p. 7
Other imperfections; Absorption
- Modeled refractive-index variations changed transmission by typically less than 1%, while nodal glia increased it by ∼2% for the thinnest axons and barely changed it for the thickest.Myelin absorption was inferred from its constituents to be negligible over ∼1 cm, but the authors note that direct measurement is needed.Source: Results · p. 7
Attainable transmission
- After 1 cm, modeled transmission is ∼31% for one 1.3 µm-wavelength example and ∼82% for a 0.61 µm-wavelength example; a thinner-axon example yields ∼3%.For the same three examples, modeled transmission after 2 mm is ∼78%, ∼96%, and ∼46%, respectively; sources and receivers should be near myelinated-section ends to reduce coupling losses.Source: Results · pp. 7–8
Attainable communication rates
- Taking an uncertain estimate of about 1 photon per neuron per minute at face value, the authors argue that aggregate emissions could still be relevant for communication.The estimate could be high because the slice was strongly stimulated or low because only scattered photons were counted; rates may vary by neuron type.Source: Results · p. 8
Proposals to test the hypothesis
- Proposed tests include coupling light into isolated living neurons in vitro, detecting light in myelin in vivo, and identifying natural neuronal sources and detectors.The authors also propose artificial sources and detectors, including fluorescent molecules or light-sensitive proteins, to test guidance in living neurons.Source: Results · p. 8
Discussion
The authors conclude that light conduction in myelinated axons is possible despite realistic imperfections and propose experiments to test the hypothesis. They also discuss possible quantum and consciousness-related implications, alongside modeling choices used to estimate transmission.
Discussion
- The authors conclude that light conduction in myelinated axons is possible despite realistic imperfections and propose experiments to test the hypothesis.They describe the optical-channel implications as conditional on optical communication along myelinated axons being real.Source: Discussion · p. 9
- Long-range quantum networking would require photons to interface with nuclear spins and those spins to interact coherently across axons.The authors mention cryptochrome-like proteins as a possible interface and synaptic junctions as potentially relevant to close contact between axons.Source: Discussion · p. 9
- Whether consciousness-related brain regions contain axons of sufficient diameter for light guidance remains an anatomical question.The authors name the claustrum, thalamus, hypothalamus, amygdala, and posterior-cortex hot zone as regions to examine.Source: Discussion · p. 9
- The simulations use FDTD and MODE software, with wavelength-dependent mesh accuracy and mode-profile checks to assess light guidance.FDTD spans a wide frequency range per simulation, while the authors also compare results against simulations using narrow sources and exact eigenmodes.Source: Discussion · p. 9
- Transmission estimates include guided power near the myelin boundaries and normalize for wavelength-dependent input-mode differences.For bends, the authors instead assess total transmitted power because the guided-mode basis changes along the continuously varying structure.Source: Discussion · p. 10
- The total-axon estimate multiplies simulated transmission fractions, exponentiating selected losses from short segments but not bend or cross-sectional-shape losses.Bend loss is treated as depending primarily on curvature change, while loss from non-circular cross-sections is treated as coupling loss.Source: Discussion · p. 11
Supplementary Information for “Possible existence of optical communication channels
The supplementary information examines transmission in myelin-guided modes, characterizing how axon structure, neighboring axons, nodal regions, and uncertain internal refractive properties affect guidance. It also defines transmission measures that distinguish guided-mode power flow from total-power transmission when cross-sections change.
Guided modes
- Simulated myelinated-axon modes were azimuthally polarized TE modes, with 82.13–99.58% of power confined in myelin across selected axon and wavelength cases.The model selected wavelengths from 0.4 µm up to the smaller of the myelin thickness and 1.3 µm, requiring at least 80% confinement; the Gaussian field profile was only an intuition, not used in simulations.Source: in the brain” · pp. 14–15, Figure S1 · p. 14, Figure S2 · p. 15
Continuously varying non-circular cross-sectional shape
- For continuously changing elliptical cross-sections, transmission depended on wavelength and aspect-ratio change; near-unity transmission occurred at ΔAR per 50 µm close to 0.40.Longer wavelengths transmitted better in the tested aspect-ratio changes, while comparisons across calibers suggested an intermediate wavelength could maximize transmission through a competition between mode availability and distortion.Source: in the brain” · pp. 15–17, Figure S3 · p. 16, Figure S3c
Cross talk between axons
- Axons should be separated by about a wavelength to avoid cross talk; larger-caliber multimode waveguides confined light better than few-mode ones at a given wavelength.For perpendicular crossings, power loss was below 0.5% across calibers; estimates that exponentiate transmission through touching axons over 50 µm are strict upper bounds on loss because power can transfer among axons.Source: in the brain” · p. 17, Figure S4 · p. 18
Nodal and paranodal regions
- With axonal guided modes included, transmission through a paranode-node-paranode region was bounded by 67.09% in all modes versus 33.78% within a wavelength of myelin for λ = 0.4 µm and r′ = 5 µm.The all-mode result is an upper bound and the myelin-span result a lower bound; across subsequent regions, transmission for most cases fluctuated around the first-pass value, supporting approximate exponentiation of single-region transmission.Source: in the brain” · pp. 18–19, Figure S5 · p. 19, Figure S6 · p. 20
Effect of the scatterers and possibility of light guidance inside the axon
- Uniformly distributed microtubules occupying about 2% of axonal volume changed transmission negligibly (± about 1%), but larger internal scatterers could strongly limit axon-guided light.For modeled larger scatterers, an axon refractive index of 1.38 gave 75.47% transmission at 0.612 µm and 95.93% at 1.3 µm, while an index of 1.35 gave 16.06% at 0.612 µm and no guided modes at 1.3 µm. Scatterers caused only a few percent variation in myelin-guided transmission, even when some output light entered the axon.Source: in the brain” · pp. 19–21
Varying the refractive indices of the axon and the cytoplasmic loops
- When the axon refractive index was 1.34 and cytoplasmic-loop index was 1.38, guided-mode transmission was higher in almost all tested cases than in the earlier simulations.The authors suggest that higher-index loops can prevent modes from diverging into the axon and can act as weak waveguides.Source: in the brain” · p. 21, Figure S7 · p. 22
in the brain”
- Integrating the real Poynting vector gives time-averaged power flow, while its imaginary part quantifies reactive power.Reactive power can arise from interference due to a standing wave.Source: in the brain” · p. 23
- After PNP regions, transmission is calculated from guided-light power across the myelin sheath, excluding a wavelength near its boundaries.Only guided light is included because nonguided light is expected to decay along an approximately uniform internode.Source: in the brain” · p. 23
- For continuously changing cross-sections, total-power transmission is more appropriate because light can shift between guided and nonguided modal bases.It is obtained by integrating the real Poynting vector for the output fields across the sheath, up to a wavelength from its boundaries.Source: in the brain” · p. 23
Supplementary Methods
The supplementary methods model changing axon cross-sections by twisting ellipses and relate that approximation to randomly shaped axons. They also specify how transmission is quantified for these shapes and for neighboring axons.
Supplementary Methods
- Twisting an elliptical axon over a 50 µm segment models continuous shape change while keeping cross-sectional area within 10% variation.The ellipse becomes circular at one-fourth length, swaps its axes halfway, and returns to its initial shape at the end.Source: Supplementary Methods · p. 23
- Ellipse and random cross-section inhomogeneity are treated as approximately equivalent, with comparable or higher loss found for the elliptical waveguide.The ellipse’s mean radial distance is within 7% of r = (a + b)/2 across simulations.Source: Supplementary Methods · p. 23
- Transmission for elliptical guides is integrated across the real Poynting flux between boundaries offset by one wavelength, with insertion and propagation losses conservatively assigned to propagation.The method normalizes larger-wavelength transmission against transmission within one wavelength of the myelin sheath for a circular guide.Source: Supplementary Methods · pp. 23–24
- Cross-talk transmission is measured as power remaining in the incident axon’s myelin sheath, normalized to transmission without the neighboring axon.Source: Supplementary Methods · p. 24