Skip to content

Axon hillock currents enable single-neuron-resolved 3D reconstruction using diamond nitrogen-vacancy magnetometry

MetadataDetails
Publication Date2020-10-02
JournalCommunications Physics
AuthorsMadhur Parashar, Kasturi Saha, Sharba Bandyopadhyay
InstitutionsIndian Institute of Technology Kharagpur, Indian Institute of Technology Bombay
Citations5
AnalysisFull AI Review Included
  • Core Value Proposition: This research validates a method for single-neuron-resolved, 3D functional brain mapping by sensing Action Potential Magnetic Fields (APMFs) using widefield diamond Nitrogen-Vacancy (NV) center magnetometry.
  • Dominant Magnetic Signature: Simulations show that intra-axonal currents in the mammalian axon hillock generate APMFs up to 36 pT (peak-to-peak), which are two orders of magnitude greater than signals from other neuronal segments.
  • Inverse Problem Simplification: The localized, dominant axon hillock signature transforms the complex 3D source reconstruction problem into detecting activity fixed in a small (~10 ”m) region, enabling reliable source localization.
  • Reconstruction Algorithm: A dictionary-based matching pursuit algorithm was developed to reconstruct spike timing and location from simulated 2D NV Magnetometric Maps (NVMM) time series.
  • Performance and Resilience: The algorithm achieved high reconstruction accuracy (up to 83.8% in 2D and 71.7% in 3D) and demonstrated exceptional resilience to Gaussian noise, requiring a minimum Signal-to-Noise Ratio (SNR) as low as -13.9 dB for reliable 2D reconstruction.
  • Spatial Resolution Achieved: The method can resolve nearby neurons separated by 20 ”m laterally (with 1 ms spike time difference) or 7 ”m axially (with 0.5 ms spike time difference), achieving single-cell resolution.
ParameterValueUnitContext
Mammalian APMF Magnitude36pTPeak-to-peak Y component magnetic field measured 20.5 ”m below the axon hillock.
APMF Signal BandwidthDC to a fewkHzFrequency range of the neuronal magnetic signal.
NVC DC-Field Sensitivity (Best Record)50pT Hz-1/2Ensemble vector magnetometry measurement record cited in the paper.
Minimum SNR (2D Lateral Case)-13.9dBRequired for reliable reconstruction (Gaussian noise, 10 ”m separation, 0.5 ms Δt).
Minimum SNR (3D Axial Case)-10.2dBRequired for reliable reconstruction (Gaussian noise, 7 ”m separation, 0.5 ms Δt).
Lateral Resolution Limit20”mMinimum separation resolved for neurons spiking with 1 ms time difference.
Axial Resolution Limit7”mMinimum separation resolved for neurons spiking with 0.5 ms time difference.
NVMM Pixel Size20 x 20”m2Spatial resolution of the simulated NVC sensor plane.
Dictionary Timepoints (ntp)3TimepointsNumber of successive timepoints (3.5, 4.0, 4.5 ms) used to form the dictionary element, focusing on peak axon hillock activity.
Pyramidal Neuron Model6250NeuronsTotal number of randomly oriented neurons simulated in the 3D volume (1 mm x 2 mm x 70 ”m).
StepDescription
1. Pyramidal Neuron SimulationUsed a realistic cortical pyramidal neuron model implemented in NEURON to simulate voltage propagation and Action Potential (AP) initiation, incorporating realistic geometry and ion-channel densities (e.g., high Na+ channel density in the Axon Initial Segment/AIS).
2. Intra-Axonal Current CalculationCalculated current flow (Icompartment) across each isopotential segment using the discrete version of the cable equation, confirming the two orders of magnitude current dominance in the axon hillock region.
3. Forward Magnetic Field ModelApplied the Biot-Savart law to calculate the vector magnetic field (Bnv) at the measurement plane by summing contributions from intra-axonal currents across all segments (Equation 5).
4. NVMM GenerationSimulated 2D NV Magnetometric Maps (NVMMs) by calculating Bnv across a 50x100 pixel grid (20 ”m pixel size) on the diamond NVC layer, generating time series data (Bt).
5. Dictionary Construction (A)Created a dictionary matrix (A) where each column (Ai) represents the concatenated 1D vector of Bx, By, and Bz components for a single neuron at three specific timepoints (3.5, 4.0, 4.5 ms) corresponding to peak axon hillock activity.
6. Matching Pursuit Algorithm (Inverse Model)Implemented a recursive dictionary-based matching pursuit algorithm (AX = B + Δ). In each iteration, the algorithm finds the neuron (k) whose dictionary element (Ak) has the maximum dot product with the experimental map (Bt) above a threshold (T).
7. Spike Detection and SubtractionA spike is confirmed if the same neuron index (k) occurs consecutively p1 times out of p2 timepoints. The NVMM signature of the detected spike is then subtracted from the experimental map (Bt), and the residual is used for the next iteration, enabling sequential spike localization.
8. Noise AnalysisTested reconstruction performance by adding both Gaussian noise (proportional to rms(S)) and shot noise (proportional to signal magnitude,

The findings support the feasibility of advanced functional brain imaging, driving demand for specialized diamond materials and quantum sensing hardware:

  • Next-Generation Functional Brain Mapping: Potential replacement for current techniques (MEG/EEG/fMRI) by offering non-invasive, high-resolution (single-cell) spatiotemporal mapping of neural activity in deep brain tissue (>1 mm depth).
  • NV Diamond Quantum Sensing: Requires high-quality, high-density NV diamond layers with optimized coherence times (T2) and high collection efficiency for widefield DC magnetometry to achieve the necessary sub-picotesla sensitivity.
  • Neuroscience Research Tools: Development of microscale magnetic imagers for in vitro studies (e.g., brain slices, organoids) to map complex neural network connectivity and dynamics at millisecond resolution.
  • Clinical Diagnostics: Enabling high-resolution localization of pathological current sources (e.g., epileptic foci) that are currently only coarsely resolved by MEG/EEG.
  • Algorithm Development: The dictionary-based matching pursuit approach provides a robust framework for solving highly ill-conditioned inverse problems in sparse source localization, applicable to other quantum sensing modalities.