Modulating Neurons With Sound — and Doing It Safely

The promise and the puzzle

Transcranial focused ultrasound (tFUS) can steer a millimetre-scale focus to a deep target through the intact skull, non-invasively, reversibly, and without genetic modification. That is a remarkable combination: it reaches DBS-class targets (for instance the thalamus) that non-invasive electrical and magnetic methods cannot focus on. The puzzle is that, at the low intensities used for modulation, we still do not fully agree on how it changes neural firing.

Candidate mechanisms

At modulatory intensities the effect is thought to be mechanical, not thermal. Leading candidates include intramembrane cavitation (the NICE model), in which the acoustic pressure flexes the lipid bilayer and changes its capacitance; direct gating of endogenous mechanosensitive channels; and steady acoustic radiation force deforming membranes. These are not mutually exclusive, and which dominates likely depends on frequency, intensity and cell type — part of why results vary across labs.

f_{\text{rad}} = \frac{2\,\alpha\, I}{c}

Acoustic radiation force density from absorption: force per unit volume scales with absorption coefficient α and intensity I. It is one of several proposed mechanical routes by which sound could perturb a membrane.

Sonogenetics: adding specificity

Just as optogenetics made light cell-type-specific, sonogenetics aims to make sound specific: express an engineered mechanosensitive channel (candidates include MscL, certain TRP channels, or the auditory protein Prestin) in target cells so that only they respond to a benign ultrasound field. The prize is optogenetic-style specificity with ultrasound's depth and skull penetration. Honest status: sonogenetics is well established in the worm and is emerging but still early in the mammalian brain, with actuator engineering and off-target effects the active problems.

The auditory confound: a cautionary tale

Safety: staying inside the window

Two dose metrics guard the safe window. The mechanical index (MI) bounds the risk of cavitation-related mechanical bioeffects, and thermal dose bounds heating. Low-intensity modulation deliberately stays well inside diagnostic limits; high-intensity ablation deliberately crosses them under imaging guidance.

\mathrm{MI} = \frac{p_{r}\,[\mathrm{MPa}]}{\sqrt{f_c\,[\mathrm{MHz}]}}

Mechanical index: peak-rarefactional (derated) pressure divided by the square root of centre frequency. Diagnostic ultrasound is capped near MI ≤ 1.9; modulation studies typically stay well below.

\mathrm{CEM43} = \sum_i t_i\, R^{\,(43 - T_i)}, \qquad R = \begin{cases} 0.25 & T < 43^{\circ}\mathrm{C} \\ 0.5 & T \ge 43^{\circ}\mathrm{C} \end{cases}

Thermal dose as cumulative equivalent minutes at 43 °C: each interval at temperature T is weighted and summed. It converts a messy temperature history into one number for safety comparison.

w_{\text{lat}} \approx \lambda\, \frac{F}{D} = \lambda \cdot F_{\#}

Diffraction-limited lateral focus width scales with wavelength times the F-number (focal length over aperture). Through the skull, aberration widens and shifts this focus, which is why a CT-informed phased array must correct the wavefront.

Finally, distinguish regimes. High-intensity focused ultrasound ablation (thalamotomy for essential tremor) is an approved clinical therapy — a destructive, not modulatory, use. Focused-ultrasound BBB opening with injected microbubbles transiently and reversibly opens the blood-brain barrier for drug delivery, and is in trials. Low-intensity reversible modulation is the youngest and least settled of the three.