Essays on Science and SocietyNeuromodulation

Nongenetic neural control with light

See allHide authors and affiliations

Science  02 Aug 2019:
Vol. 365, Issue 6452, pp. 457
DOI: 10.1126/science.aay4351

Extracellular electrical stimulation of neurons forms the basis for many implantable disease-treating devices. However, these devices—usually metal electrodes—are often limited by their bulk, mechanical invasiveness, and inability to target individual neurons and neuronal circuits.

A more precise alternative to implantable devices, optogenetics, uses light to selectively control neurons. However, optogenetics requires genetic modifications, which means that using it as a treatment in humans is challenging. As a result, neurostimulation treatment is limited to patients with severe and complex neurological conditions, and the mechanisms by which neurostimulation improves symptoms (such as reduction of seizures) remain poorly understood. My research seeks to find a nongenetic solution to this problem that will allow minimally invasive neuromodulation at the single-cell or subcellular level with the flexibility of optical stimulation.

Embedded Image

My academic path has exposed me to photovoltaics, electrophysiology, and tissue engineering, all of which have led me to appreciate that the nature of signal transduction and communication at neuron-device interfaces is essentially one of energy exchanges. As such, my research explores ideas from energy sciences and applies them to neuromodulation.

My lab uses silicon-based nanostructured materials to understand the fundamental bioelectric dynamics of individual cells, organelles, and their networks. Building on this foundation, we have designed a neuromodulation method based on the optical properties of silicon. It usually takes freestanding forms and can be operated with high flexibility and spatial resolution to implement multiplexed and patterned neural modulation.

The Concept

Silicon nanostructures are biocompatible materials that can be used to target a single cell or subcellular component. Silicon has highly tunable electrical and optical properties and can absorb a broad range of light wavelengths.

We already have a deep understanding of silicon's behavior in energy sciences. When silicon absorbs light, the resulting electronic energy can be converted into heat (i.e., photothermal process) or chemical energy (i.e., photoelectrochemical process). At neuron-device interfaces, this photothermal or photoelectrochemical process depolarizes the plasma membrane, which elicits an action potential in the cells. Essentially, our method for optical extracellular neuromodulation is a nongenetic version of the technique known as “optogenetics.”

The Potential

Freestanding, wireless, silicon-based biomedical devices for neurostimulation can easily interface with biological systems. They have the potential to overcome the limitations of current metal electrode–based devices, which include bulk and cell membrane disruption, and are not dependent on genetic modifications. Moreover, the ability of silicon to absorb light in the infrared regime can be useful for penetrating tissue.

Research in my lab has revealed how the energy outputs from the photothermal and photoelectrochemical effects of nanostructured silicon can be identified, quantified, and used at silicon-based bio interfaces to modulate electrical activities in neurons.

Photothermal Modulation of Neurons

In 2016, my lab synthesized a deformable and porous type of silicon with molecularlevel feature sizes (1). The measured Young's modulus at the material surface is ∼0.41 GPa in phosphate-buffered saline (PBS) solution, which is comparable to that of hydrated collagen fibers. The deformable silicon shows enhanced light absorption, as well as reduced thermal conductivity and heat capacity, all of which result in a rapid photothermal heating. We took this opportunity to explore how silicon's photothermal effect could be used to modulate signal transduction at a neuron–silicon interface (1) (see the figure).

We first placed the deformable silicon on phospholipid bilayers, the fundamental building blocks of cell membranes. Illumination of the lipid-supported silicon particles yielded a fast photothermal effect in the silicon, which in turn induced a local temperature increase. This fast temperature increase produced a transient capacitive current from the phospholipid bilayer and bilayer depolarization.

We then asked if this photothermally induced electric effect could be applied to living cell systems (1). We introduced the silicon particles to dorsal root ganglia (DRG) neuron cultures and illuminated the cell-membrane–supported particles, eliciting action potentials in individual neurons. We also successfully delivered a train of light pulses and repeatedly excited neurons with a one-pulse-one-spike fidelity. This confirmed that the photothermally induced electric effect could indeed be applied to living neurons.

Photothermal neuromodulation with deformable silicon uses phonons of an indirect bandgap semiconductor. Silicon's photothermal effect at the neuron–silicon interface does not require direct physical contact because the heating can be effective for a distance up to 100 µm. This makes it ideal for use in situations such as peripheral nerve stimulations where extracellular matrix or other cellular barriers would usually impede tight biointerfaces.

Photoelectrochemical Modulation of Neurons

For greater efficiency in neuromodulation, a tight interface between the silicon device and the neuron is required. When direct access to the cells is available, the preferred neuromodulation approach would be to use electrons and holes (i.e., the charge carriers) that are generated by light —the same way a photoelectrochemical device works.

Silicon nanowire–based treatment is an attractive option for neurostimulation. Because of their small size, silicon nanowires can be delivered into neural cultures or tissues in a drug-like manner with high spatial resolution. The silicon surfaces can be modified with peptides or other macromolecules to allow for high-affinity binding to specific neural cells.

To investigate these capabilities, my lab used coaxial p-type/intrinsic/n-type silicon (PINS) nanowires to modulate primary rat DRG neuron excitability (2) (see the figure). We proposed that upon light stimulation at a neuron–silicon nanowire interface, electrons move toward the n-type shell and holes move to the p-type core; this induces a cathodic photoelectrochemical process at the n-shell that can reduce the local extracellular potential for neuron depolarization.

Silicon-based nongenetic optical neuromodulation

(A and B) Rapid photothermal process from deformable silicon (A) excites neural activities (B) by membrane capacitance changes. (C and D) Photoelectrochemical neuromodulation. Schematic of the photoelectrochemical reaction at neural interfaces eliciting action potentials (C) and an optical excitation curve (D). (E and F) Our rational design principle for neuromodulation interfaces allows us to control neuromodulation at multiple biological length scales, such as control of limb motion from anesthetized mice.


The nanowires that we used had been synthesized during my graduate studies in 2007 (3). When we reexamined them, we discovered the presence of atomic gold on the nanowire surface, likely due to gold migration during silicon deposition. We wondered whether the surface gold played a role in the photoelectrochemical current generation and localized neuromodulation induced by single nanowires.

We used a quartz microelectrode and recorded sustained light-triggered electrochemical currents from the surface of single nanowires. Our results showed that the atomic gold on the nanowires enhances the photoelectrochemical process through which the action potentials in rat DRG neurons were elicited (2). Atomic gold reduces the kinetic barrier for photoelectrochemical current generation, essentially playing the role of a catalyst in traditional photoelectrochemical devices.

Formulation of Design Principles

Our quest to establish a rationally designed and coherent framework for semiconductor-based neuromodulation tools (4) continues.

Efficient modulation requires accurate designs for tight neuron–device interfaces. We recently identified a biology-guided two-step design principle for establishing tight intra-, inter-, and extracellular silicon-based biointerfaces in which silicon and neural targets have matched mechanical properties and efficient signal transduction. We showed nongenetic optical modulation capabilities across multiple biological length scales. Specifically, we demonstrated that silicon-based materials (4) in the forms of a deformable and distributed mesh (organ level), a multilayered membrane with nanostructured surfaces (cellular and tissue levels), and a nanocrystalline nanowire (organelle level) promote tight neural interfaces.

Neurons receive, process, and communicate messages in different ways. To gain a biophysical understanding of the different biological modulations that silicon could induce, my lab developed several matrices to quantify the light-triggered photothermal and photoelectrochemical outputs from ∼30 different silicon materials in PBS (4). We confirmed that we could use light to achieve nongenetic modulations of subcellular calcium concentrations, cytoskeletal organization and transport, and neuronal excitability. In particular, we were able to show that the deformable and distributed silicon mesh can enable light-controlled modulation of brain activities and simple animal behaviors such as induced limb motion from anesthetized mice (4) (see the figure).

What's Next?

By adopting the science and technical approaches of energy research, our lab has developed a new semiconductor-based platform (1, 2, 46) for remotely controlled nongenetic biological modulations (see the figure), including the optical modulation of neurons and heart cells. The minimally invasive integration of semiconductor-based neuromodulation devices with the nervous system will help to address unanswered questions in neuroscience and neuroengineering, reevaluate some principles that were believed to be understood, and aid in the implementation of new therapeutics for neurological disorders.

Embedded Image


Bozhi Tian received his undergraduate and master's degrees from Fudan University and his Ph.D. from Harvard University. After postdoctoral training at Massachusetts Institute of Technology and Harvard Medical School, Tian started his lab in the Department of Chemistry at the University of Chicago in 2012. His current research involves bioelectronics, semiconductorenabled probing of subcellular biophysics, and the chemical dynamics at soft-hard interfaces.

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article