Octopus visual system: a functional MRI model for detecting neuronal electric currents without a blood-oxygen-level-dependent confound

Can MRI Be Used as a Sensor to Record Neural Activity?

Magnetic resonance provides exquisite anatomical images and functional MRI monitors physiological activity by recording blood oxygenation. This review attempts to answer the following question: Can MRI be used as a sensor to directly record neural behavior? It considers MRI sensing of electrical activity in the heart and in peripheral nerves before turning to the central topic: recording of brain activity. The primary hypothesis is that bioelectric current produced by a nerve or muscle creates a magnetic field that influences the magnetic resonance signal, although other mechanisms for detection are also considered. Recent studies have provided evidence that using MRI to sense neural activity is possible under ideal conditions. Whether it can be used routinely to provide functional information about brain processes in people remains an open question. The review concludes with a survey of artificial intelligence techniques that have been applied to functional MRI and may be appropriate for MRI sensing of neural activity.

In vivo direct imaging of neuronal activity at high temporospatial resolution

There has been a long-standing demand for noninvasive neuroimaging methods that can detect neuronal activity at both high temporal and high spatial resolution. We present a two-dimensional fast line-scan approach that enables direct imaging of neuronal activity with millisecond precision while retaining the high spatial resolution of magnetic resonance imaging (MRI). This approach was demonstrated through in vivo mouse brain imaging at 9.4 tesla during electrical whisker-pad stimulation. In vivo spike recording and optogenetics confirmed the high correlation of the observed MRI signal with neural activity. It also captured the sequential and laminar-specific propagation of neuronal activity along the thalamocortical pathway. This high-resolution, direct imaging of neuronal activity will open up new avenues in brain science by providing a deeper understanding of the brain's functional organization, including the temporospatial dynamics of neural networks.

Diffusion MRI Connections in the Octopus Brain

Using high angle resolution diffusion magnetic resonance imaging (HARDI) with fiber tractography analysis we map out a meso-scale connectome of the Octopus bimaculoides brain. The brain of this cephalopod has a qualitatively different organization than that of vertebrates, yet it exhibits complex behavior, an elaborate sensory system and high cognitive abilities. Over the last 60 years wide ranging and detailed studies of octopus brain anatomy have been undertaken, including classical histological sectioning/staining, electron microscopy and neuronal tract tracing with injected dyes. These studies have elucidated many neuronal connections within and among anatomical structures. Diffusion MRI based tractography utilizes a qualitatively different method of tracing connections within the brain and offers facile three-dimensional images of anatomy and connections that can be quantitatively analyzed. Twenty-five separate lobes of the brain were segmented in the 3D MR images of each of three samples, including all five sub-structures in the vertical lobe. These parcellations were used to assay fiber tracings between lobes. The connectivity matrix constructed from diffusion MRI data was largely in agreement with that assembled from earlier studies. The one major difference was that connections between the vertical lobe and more basal supra-esophageal structures present in the literature were not found by MRI. In all, 92 connections between the 25 different lobes were noted by diffusion MRI: 53 between supra-esophageal lobes and 26 between the optic lobes and other structures. These represent the beginnings of a mesoscale connectome of the octopus brain.

Estimation of the minimum detectable phase change of surface coil for neural current MRI

Neuronal current magnetic resonance imaging (NC-MRI) is a new method in functional imaging of the brain that could cause the alteration in the phase of magnetic resonance signal. The phase variance is defined as the inverse of the signal to noise ratio (SNR). The intrinsic SNR of the MRI signal is characterized by the coil performance. We evaluated the relation between the geometry and the shape of coils in order to find the minimum detectable change in the signal phase and the possibility of direct detection of neuronal activity by MRI. Full wave equations were solved by the finite element method to calculate the SNR for circular, elliptical, and square shape surface coils. The simulation was repeated for Larmor frequencies of 64 MHz and 85.2 MHz and the coil sizes between 1.5 and 7.5 cm. Relative intrinsic signal to noise ratio (rISNR) of coils with a respect to a selected reference coil and a reference point in the sample was estimated. The circular coil had higher rISNR than other shapes. The increase of the strip width in the coils raised the rISNR 5-20%. For typical imaging parameters, rISNR reference was about 66 which led to a minimum detectable change in MRI signal phase of 0.87° (11.4 nT). It may also be reduced up to tenfold in a 1.5 cm circular coil. Detection of subtle phase signal change due to neuronal activity in surface coils needs a large amount of data acquisition and averaging, but it is intrinsically feasible.

Functional magnetic resonance electrical impedance tomography (fMREIT) sensitivity analysis using an active bidomain finite-element model of neural tissue

PURPOSE

A direct method of imaging neural activity was simulated to determine typical signal sizes.

METHODS

An active bidomain finite-element model was used to estimate approximate perturbations in MR phase data as a result of neural tissue activity, and when an external MR electrical impedance tomography imaging current was added to the region containing neural current sources.

RESULTS

Modeling-predicted, activity-related conductivity changes should produce measurable differential phase signals in practical MR electrical impedance tomography experiments conducted at moderate resolution at noise levels typical of high field systems. The primary dependence of MR electrical impedance tomography phase contrast on membrane conductivity changes, and not source strength, was demonstrated.

CONCLUSION

Because the injected imaging current may also affect the level of activity in the tissue of interest, this technique can be used synergistically with neuromodulation techniques such as deep brain stimulation, to examine mechanisms of action.

Direct detection of neural activity in vitro using magnetic resonance electrical impedance tomography (MREIT)

We describe a sequence of experiments performed in vitro to verify the existence of a new magnetic resonance imaging contrast - Magnetic Resonance Electrical Impedance Tomography (MREIT) -sensitive to changes in active membrane conductivity. We compared standard deviations in MREIT phase data from spontaneously active Aplysia abdominal ganglia in an artificial seawater background solution (ASW) with those found after treatment with an excitotoxic solution (KCl). We found significant increases in MREIT treatment cases, compared to control ganglia subject to extra ASW. This distinction was not found in phase images from the same ganglia using no imaging current. Further, significance and effect size depended on the amplitude of MREIT imaging current used. We conclude that our observations were linked to changes in cell conductivity caused by activity. Functional MREIT may have promise as a more direct method of functional neuroimaging than existing methods that image correlates of blood flow such as BOLD fMRI.

Estimation of phase signal change in neuronal current MRI for evoke response of tactile detection with realistic somatosensory laminar network model

Magnetic field generated by neuronal activity could alter magnetic resonance imaging (MRI) signals but detection of such signal is under debate. Previous researches proposed that magnitude signal change is below current detectable level, but phase signal change (PSC) may be measurable with current MRI systems. Optimal imaging parameters like echo time, voxel size and external field direction, could increase the probability of detection of this small signal change. We simulate a voxel of cortical column to determine effect of such parameters on PSC signal. We extended a laminar network model for somatosensory cortex to find neuronal current in each segment of pyramidal neurons (PN). 60,000 PNs of simulated network were positioned randomly in a voxel. Biot-savart law applied to calculate neuronal magnetic field and additional phase. The procedure repeated for eleven neuronal arrangements in the voxel. PSC signal variation with the echo time and voxel size was assessed. The simulated results show that PSC signal increases with echo time, especially 100/80 ms after stimulus for gradient echo/spin echo sequence. It can be up to 0.1 mrad for echo time = 175 ms and voxel size = 1.48 × 1.48 × 2.18 mm(3). With echo time less than 25 ms after stimulus, it was just acquired effects of physiological noise on PSC signal. The absolute value of the signal increased with decrease of voxel size, but its components had complex variation. External field orthogonal to local surface of cortex maximizes the signal. Expected PSC signal for tactile detection in the somatosensory cortex increase with echo time and have no oscillation.

A comprehensive study of sensitivity in measuring oscillatory magnetic fields using rotary saturation pulse sequences

Detecting the oscillatory currents with a specific frequency distribution may have the potential to make neuronal current MRI (ncMRI) come true. The phase shift or dephasing induced by both positive and negative episodes of oscillatory neuronal currents is likely to be canceled out over the echo time in typical BOLD-contrast fMRI experiments. Based on the contrast of rotary saturation, both of the recently developed spin-locked oscillatory excitation (SLOE) and stimulus-induced rotary saturation (SIRS) pulse sequences have been demonstrated to be able to detect weak oscillatory magnetic fields in phantoms with 3T MR scanners. In this report, through Bloch equation simulation as well as water phantom and anesthetic rats experiments, we comprehensively evaluate and compare the sensitivities of these two methods (SLOE and SIRS) in detecting the oscillatory magnetic fields for both high (100 Hz) and low (10 Hz) oscillation frequencies, while using their respective optimal imaging parameters. In agreement with the theoretical predications, both the simulated and experimental results showed that the SLOE method features a much higher detection sensitivity of weak magnetic fields than that of the SIRS method. SLOE was able to detect applied oscillatory magnetic fields as low as 0.1 nT in a water phantom and 0.5 nT in rat brains and the deteriorated noise levels in rat data may account for the reduced sensitivity in vivo. These promising results form the foundation for direct detection of in vivo neuronal currents using MRI.

Computational Modeling of Neuronal Current MRI Signals with Rat Somatosensory Cortical Neurons

Magnetic field generated by active neurons has recently been considered to determine location of neuronal activity directly with magnetic resonance imaging (MRI), but controversial results have been reported about detection of such small magnetic fields. In this study, multiple neuronal morphologies of rat tissue were modeled to investigate better estimation of MRI signal change produced by neuronal magnetic field (NMF). Ten pyramidal neurons from layer II to VI of rat somatosensory area with realistic morphology, biophysics, and neuronal density were modeled to simulate NMF of neuronal tissue, from which effects of NMF on MRI signals were obtained. Neuronal current MRI signals, which consist of relative magnitude signal change (RMSC) and phase signal change (PSC), were at least three and one orders of magnitude less than a tissue with single neuron type, respectively. Also, a reduction in voxel size could increase signal alterations. Furthermore, with selection of zenith angle of external main magnetic field related to tissue surface near to 90°, RMSC could be maximized. This value for PSC would be 90° for small voxel size and zero degree for large ones.