Current State of Potential Mechanisms Supporting Low Intensity Focused Ultrasound for Neuromodulation

Abstract

Low intensity focused ultrasound (LIFU) has been gaining traction as a non-invasive neuromodulation technology due to its superior spatial specificity relative to transcranial electrical/magnetic stimulation. Despite a growing literature of LIFU-induced behavioral modifications, the mechanisms of action supporting LIFU's parameter-dependent excitatory and suppressive effects are not fully understood. This review provides a comprehensive introduction to the underlying mechanics of both acoustic energy and neuronal membranes, defining the primary variables for a subsequent review of the field's proposed mechanisms supporting LIFU's neuromodulatory effects. An exhaustive review of the empirical literature was also conducted and studies were grouped based on the sonication parameters used and behavioral effects observed, with the goal of linking empirical findings to the proposed theoretical mechanisms and evaluating which model best fits the existing data. A neuronal intramembrane cavitation excitation model, which accounts for differential effects as a function of cell-type, emerged as a possible explanation for the range of excitatory effects found in the literature. The suppressive and other findings need additional theoretical mechanisms and these theoretical mechanisms need to have established relationships to sonication parameters.

Keywords: focused ultrasound stimulation; low intensity focused ultrasound; neuromodulation; non-invasive brain stimulation; transcranial focused ultrasound.

Figure 1

Low intensity focused ultrasound general principles. (A) A depiction of a typical LIFU experimental setup. A participant is seated (2) with an US device (5) firmly pressed against their head held in place by an arm (3). The US device is controlled by a computer system (4) and targeted using infrared system (1). (B) Depiction of the mechanical wave properties (amplitude, wavelength, and frequency) used in US stimulation. (C) Spatial intensities of the mechanical wave. (D) Temporal intensities of the mechanical wave. (E) Two exemplary pulsation schemes: pulsed (in yellow) and continuous (in teal). Both the pulsing schemes have a customizable sonication duration with inter stimulation interval with the DC parameter (i.e., the ratio of tone burst duration over pulse repetition period) determining the pulsing scheme.

Figure 2

Proposed ultrasonic stimulation's mechanisms for neuromodulation. Depicted in column 1 are six neuronal membranes (four with an ion channel [rows A,C,D,E] and two neuronal membranes [rows B,F] with polar lipid bilayer) and a neuron with the microtubules highlighted (row G). Depicted in column 2, these membranes have four types of electrophysiological-mechanical coupling during an action potential: change in membrane conformation state, thermodynamic waves, direct flexoelectricity, and opening of mechanosensitive ion channels (see Section above). Column 3 depicts these same four electrophysiological-mechanical coupling during US stimulation along with three other possible mechanisms of US's neuromodulation: thermal modulation, sonoporation and cavitation, and microtubule resonance (see Section above).

Figure 3

Neuronal intramembrane cavitation excitation model. Plaksin et al. (2014, 2016) proposed the NICE model hypothesizing sonoporation (see Section above and Figure 2F) as US's mechanism of neuromodulation. The US's DC (see Figure 1E) determines the polarity of neuromodulation. A low DC (i.e., below 5%) during a stimulation's off-periods will preferentially activate thalamic reticular neurons (TRN), thalamocortical neurons (TCN), and low-threshold spiking (LTS) interneurons via T-type voltage-gated calcium channels (see Section above for full description) producing an inhibitory effect. A high DC (i.e., over 20%) during the on-periods will preferentially activate regular spiking (RS) pyramidal cells and fast spiking (FS) interneurons while suppressing the LTS interneurons producing an overall excitatory effect. This excitatory effect is simulated using a basic network model of LTS, FS, and RS neurons connected with excitatory and inhibitory synapses and thalamic inputs. The network model predicts an optimum excitation of 70% DC.

Figure 4

Excitatory and suppressive empirical findings' relationships to DC, PRF, I_SPPA, f_c, and SD. DC, PRF, I_SPPA, f_c, and SD are used as grouping factors for excitatory and suppressive findings. We used density plots for each study, but studies with multiple sonication parameters have each one plotted. In the top panel, high DC, above 10%, has the vast majority of the excitatory findings. While ow DC, less than 10%, contains the majority of the suppressive findings, there are still approximately 30% of the suppressive findings above 10% DC. The top panel is highlighted in red because DC is the one sonication parameter that has any distinction between excitatory and suppressive findings. In the four bottom panels, PRF, I_SPPA, f_c, and SD has no clear distinction between excitatory and suppressive findings.