Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 86 of the 89 references it cites were found in a registry or an index; none of the 52 sources its claims cite is checked for support yet.

High-fidelity control of spin ensemble dynamics via artificial intelligence: from quantum computing to NMR spectroscopy and imaging

Open access

Abstract

High-fidelity control of spin ensemble dynamics is essential for many research areas, spanning from quantum computing and radio-frequency (RF) engineering to NMR spectroscopy and imaging. However, attaining robust and high-fidelity spin operations remains an unmet challenge. Using an evolutionary algorithm and artificial intelligence (AI), we designed new RF pulses with customizable spatial or temporal field inhomogeneity compensation. Compared with the standard RF shapes, the new AI-generated pulses show superior performance for bandwidth, robustness, and tolerance to field imperfections. As a benchmark, we constructed a spin entanglement operator for the weakly coupled two-spin-1/2 system of 13CHCl3, achieving high-fidelity transformations under multiple inhomogeneity sources. We then generated band-selective and ultra-broadband RF pulses typical of biomolecular NMR spectroscopy. When implemented in multipulse NMR experiments, the AI-generated pulses significantly increased the sensitivity of medium-size and large protein spectra relative to standard pulse sequences. Finally, we applied the new pulses to typical imaging experiments, showing a remarkable tolerance to changes in the RF field. These AI-generated RF pulses can be directly implemented in quantum information, NMR spectroscopy of biomolecules, magnetic resonance imaging techniques for in vivo and materials sciences.

Overview

CoChat Summary

TL;DR

The paper introduces GENETICS-AI, which combines an evolutionary algorithm and a neural network to design constant-amplitude RF phase pulses with high fidelity and compensation for field inhomogeneities. It demonstrates these AI-generated pulses for spin entanglement, biomolecular NMR spectroscopy, and MRI, reporting improved bandwidth, robustness, and sensitivity over standard rectangular pulses.

Key findings

  • GENETICS-AI generated RF pulses with operational fidelity up to 0.99999 and a library of ∼200,000 phase shapes; for arbitrary flip angles, broadband pulses reached fidelity >0.9999, including a 113.5° pulse with average fidelity 0.9999 over 50 kHz at constant 25 kHz amplitude.
  • For the weakly coupled two-spin-1/2 system of 13CHCl3, the GEN-Entangler gave a fidelity volume approximately 100 times larger than a rectangular-pulse C-NOT entangler at maximum inhomogeneity; at a 0.98 fidelity iso-surface, the volume ratio VGEN/VRECT increased to 200 times, and the sequence was ∼62% shorter than a ΔJ-compensated C-NOT gate sequence.
  • In NMR applications, universal π/2 and π pulses covered 40 kHz bandwidth with lengths <300 µs, constant ω1 = 25 kHz, and average fidelity 0.9999; ultra-broadband π/2 and π pulses reached 30 × ω1 = 750 kHz at ω1 = 25 kHz in ∼1.5 ms with 2,000 points, and an inversion pulse achieved fidelity 0.99999 at 274 µs, 100 kHz bandwidth, and 25 kHz RF amplitude. GEN-HSQC improved S/N and detected additional amide peaks in 42 kDa PKA-C, tolerating ±20% RF inhomogeneity, whereas PFG-HSQC intensities dropped from ∼0.8 near the 1H carrier to ∼0.5 off-resonance and were barely observable at ±20% attenuation.
  • In MRI, the GEN-SE sequence produced high-quality 1D phantom images with fewer artifacts at ω1 inhomogeneity up to ±80%, while standard rectangular pulses degraded significantly.

Methods

  • Design: GENETICS-AI combines iterative forward search and self-correcting algorithms (IFSA/ISCA) that search RF phase space to generate an Optimal Phase Surface library, then a neural network (PhaseNET) trained on this library predicts new RF shapes. The fitness function is average operational fidelity over offset and RF amplitude. The software was written in MATLAB R2018a and run on an Intel Core i7-7700 K; a 100-point OPS took ∼12 hours, a 500-point shape ∼1 week, and up to 2,000-point shapes were implemented, generating >200,000 RF shapes.
  • Experiments: NMR experiments used a Bruker 600 MHz spectrometer for 13CHCl3 (temperature reported as 300 K in the Figure 1 caption and 290 K in the Experimental section), a Bruker 850 MHz spectrometer for U-15N UBIK48C at 300 K and 500 µM, PKA-C (42 kDa) at 273 K and 200 µM, and 13C-labeled MBP; MRI phantom experiments used a Bruker 900 MHz spectrometer. Comparisons used standard rectangular, PFG-HSQC, and spin-echo sequences.
  • Analysis: simulated fidelity iso-surfaces and offset/RF responses, quantum state tomography, 2D HSQC spectra, and 1D imaging.

Limitations

  • Stated: the algorithm can currently design only pulses with constant amplitude or preselected amplitude shapes; future work will include coding RF pulse amplitude in the search space. The 5 GHz biomolecular NMR pulses are hypothetical because commercially available ultra-high magnetic fields are limited to 1.2 GHz. The ultra-broadband pulse showed signal loss due to relaxation during pulse execution. Computational cost grows exponentially (100-point OPS ∼12 hours; 500-point shape ∼1 week), and digital resolution limits the maximum achievable bandwidth.
  • Evident from the text: MRI validation is a 1D phantom experiment, not in vivo imaging, and the NMR demonstrations are proof-of-concept applications on selected proteins and samples.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1093/pnasnexus/pgac133

Cites

86 of 86 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1. Generation of spin entanglement for 13CHCl3. (A) 13CHCl3 used for validating the GEN-Entangler pulse sequence. (B) GEN-Entangler pulse sequence. All the pulses are global for both 1H and 13C nuclei. Pulses and delay parameters for different values of n are given in Table S1. (C) J response curve of entanglement fidelity for n = 1, 2, and 4. (D) and (E) RF shapes and corresponding fidelity profiles used for GEN-Entangler with n = 2. (F) Quantum state tomography for the real and imaginary part of the }{}$|{\rm{\alpha \alpha }}\rangle $ pseudo pure state created for 1H and 13C nuclear spins. (G) Simulated iso-surface at 0.99 fidelity for the entanglement created with n = 2 (outer surface—yellow) and standard C-NOT entangler with rectangular-shaped pulse (inner surface—red). The quantum state tomography results for simulated inhomogeneities on the 0.99 fidelity iso-surface for both extreme values of all three variables are shown in blue. (H) Fidelity volume fraction of entanglement created with GEN-Entangler (VGEN) over standard C-NOT with rectangular pulses (VRECT) for fidelities of 0.995, 0.99, and 0.98. All experiments were performed on a Bruker 600 MHz spectrometer at 300 K.
Figure 1. Generation of spin entanglement for 13CHCl3. (A) 13CHCl3 used for validating the GEN-Entangler pulse sequence. (B) GEN-Entangler pulse sequence. All the pulses are global for both 1H and 13C nuclei. Pulses and delay parameters for different values of n are given in Table S1. (C) J response curve of entanglement fidelity for n = 1, 2, and 4. (D) and (E) RF shapes and corresponding fidelity profiles used for GEN-Entangler with n = 2. (F) Quantum state tomography for the real and imaginary part of the }{}$|{\rm{\alpha \alpha }}\rangle $ pseudo pure state created for 1H and 13C nuclear spins. (G) Simulated iso-surface at 0.99 fidelity for the entanglement created with n = 2 (outer surface—yellow) and standard C-NOT entangler with rectangular-shaped pulse (inner surface—red). The quantum state tomography results for simulated inhomogeneities on the 0.99 fidelity iso-surface for both extreme values of all three variables are shown in blue. (H) Fidelity volume fraction of entanglement created with GEN-Entangler (VGEN) over standard C-NOT with rectangular pulses (VRECT) for fidelities of 0.995, 0.99, and 0.98. All experiments were performed on a Bruker 600 MHz spectrometer at 300 K.CC BY-NC-ND · PubMed Central
Figure 2. BB high-fidelity arbitrary flip angle pulses generated with GENETICS-AI. (A) OPS of a BB arbitrary state preparation with fidelity > 0.99999. The Y-axis represent the zenith angle (ζ) of the final state. The OPS was generated with an initial state fixed at }{}$|\alpha \rangle $ and varying the ζ from 0 to π iteratively. The total nutation angle of the pulse was fixed at 8.68 π for all states. (B) OPS for arbitrary flipping operation with fidelity > 0.9999. The Y-axis represents the flip angle (θ) of the pulse. The OPS was created with θ as a target variable. The total nutation angle for the operation was fixed at 11.42 π. (C) Bloch sphere trajectories for the preparation of four different states (A, B, C, and D) using RF shapes from the OPS shown in (A). (D) Bloch sphere trajectories of two different flipping operations from A and A’ to B and B’ using RF shape from the OPS shown in (B). All trajectories are trimmed to show only 10% of the initial and final part. (E) RF shape of 113.5° universal flipping for a bandwidth of 50 kHz. The RF amplitude is constant (25 kHz). (F) 2D-fidelity response with offset and RF amplitude of the RF shape shown in (E).
Figure 2. BB high-fidelity arbitrary flip angle pulses generated with GENETICS-AI. (A) OPS of a BB arbitrary state preparation with fidelity > 0.99999. The Y-axis represent the zenith angle (ζ) of the final state. The OPS was generated with an initial state fixed at }{}$|\alpha \rangle $ and varying the ζ from 0 to π iteratively. The total nutation angle of the pulse was fixed at 8.68 π for all states. (B) OPS for arbitrary flipping operation with fidelity > 0.9999. The Y-axis represents the flip angle (θ) of the pulse. The OPS was created with θ as a target variable. The total nutation angle for the operation was fixed at 11.42 π. (C) Bloch sphere trajectories for the preparation of four different states (A, B, C, and D) using RF shapes from the OPS shown in (A). (D) Bloch sphere trajectories of two different flipping operations from A and A’ to B and B’ using RF shape from the OPS shown in (B). All trajectories are trimmed to show only 10% of the initial and final part. (E) RF shape of 113.5° universal flipping for a bandwidth of 50 kHz. The RF amplitude is constant (25 kHz). (F) 2D-fidelity response with offset and RF amplitude of the RF shape shown in (E).CC BY-NC-ND · PubMed Central
Figure 3. BB universal pulses and comparison of GENETICS-AI refocusing pulses with previously reported shapes. (A) Plot of the “Refocusing bandwidth-pulse length” for best performing GENETICS-AI pulses versus reported refocusing pulses. Refocusing BW of each pulse is calculated by averaging the offset responses of Mx and My magnetizations with a cutoff operational fidelity of 0.95. Each pulse is labeled with a short form and color-coded for the same family. Supplementary Table S2 shows the full names and relative reference for each pulse analyzed. The size of each dot represents the average operational fidelity of the pulse. Offset responses of a few selected RF shapes are shown in Figure S9. (B) Ultra-BB universal π/2 and (C), π pulse of BW 30 × ω1. Amplitude shapes of these pulses are constant (ω1) with phase shapes shown on the left. Offset response of π and π/2 are shown for initial states Mz (middle). Simulated fidelity response with offset and RF amplitude are shown on the right panel.
Figure 3. BB universal pulses and comparison of GENETICS-AI refocusing pulses with previously reported shapes. (A) Plot of the “Refocusing bandwidth-pulse length” for best performing GENETICS-AI pulses versus reported refocusing pulses. Refocusing BW of each pulse is calculated by averaging the offset responses of Mx and My magnetizations with a cutoff operational fidelity of 0.95. Each pulse is labeled with a short form and color-coded for the same family. Supplementary Table S2 shows the full names and relative reference for each pulse analyzed. The size of each dot represents the average operational fidelity of the pulse. Offset responses of a few selected RF shapes are shown in Figure S9. (B) Ultra-BB universal π/2 and (C), π pulse of BW 30 × ω1. Amplitude shapes of these pulses are constant (ω1) with phase shapes shown on the left. Offset response of π and π/2 are shown for initial states Mz (middle). Simulated fidelity response with offset and RF amplitude are shown on the right panel.CC BY-NC-ND · PubMed Central
Figure 4. RF inhomogeneity compensated high-resolution multipulse NMR experiment. (A) GENETICS-AI shapes for universal π and π/2 pulses for 1H (blue) and 15N (red) channels. RF amplitudes were constant at 12.5 and 5.8 kHz for 1H and 15N channels, respectively. The pulse lengths for π/2 and π pulses were 488.6 and 509.8 µs for 1H and 483.8 and 524 µs for 15N. (B) Simulated fidelity response with offset and RF amplitude for 1H (blue) and 15N (red) for the pulse shapes in (A). The inner contour level indicates a fidelity of 0.99. (C) GENETICS-AI version of the HSQC (GEN-HSQC) spectrum of 15N labeled K48C mutant of Ubiquitin recorded using calibrated amplitude values for 1H (12.5 kHz) and 15N (5.8 kHz) channels on a Bruker 850 MHz spectrometer at 300 K. The insets show the intensity comparison of the standard PFG-HSQC (using the Bruker pulse sequence—“hsqcetfpf3gpsi2”) (}{}${I}_{RECT}$) and GEN-HSQC (}{}${I}_{GEN}$) at different RF amplitudes for selected residues. Left Y-axis shows the normalized peak intensities of standard (blue) and GEN-HSQC (red) relative to the calibrated reference spectra (}{}$I_{RECT}^0$ and }{}$I_{GEN}^0$), whereas the right Y-axis represents the normalized intensity of standard HSQCs (blue) with respect to }{}$I_{GEN}^0$. (D) GEN-HSQC spectra of 200 µM 15N labeled catalytic subunit of protein kinase A (42 kDa). The figure insets compare the intensity of various resonances recorded using GEN-HSQC (red) and PFG-HSQC (blue) of the HSQC experiment. Both experiments are recorded on a Bruker 850 MHz spectrometer at 273 K using 128 complex indirect points were acquired with 16 scans per FID and a 1.5 s relaxation delay. Maximum RF amplitudes were 16.67 kHz for 1H and 7.7 kHz for 15N.
Figure 4. RF inhomogeneity compensated high-resolution multipulse NMR experiment. (A) GENETICS-AI shapes for universal π and π/2 pulses for 1H (blue) and 15N (red) channels. RF amplitudes were constant at 12.5 and 5.8 kHz for 1H and 15N channels, respectively. The pulse lengths for π/2 and π pulses were 488.6 and 509.8 µs for 1H and 483.8 and 524 µs for 15N. (B) Simulated fidelity response with offset and RF amplitude for 1H (blue) and 15N (red) for the pulse shapes in (A). The inner contour level indicates a fidelity of 0.99. (C) GENETICS-AI version of the HSQC (GEN-HSQC) spectrum of 15N labeled K48C mutant of Ubiquitin recorded using calibrated amplitude values for 1H (12.5 kHz) and 15N (5.8 kHz) channels on a Bruker 850 MHz spectrometer at 300 K. The insets show the intensity comparison of the standard PFG-HSQC (using the Bruker pulse sequence—“hsqcetfpf3gpsi2”) (}{}${I}_{RECT}$) and GEN-HSQC (}{}${I}_{GEN}$) at different RF amplitudes for selected residues. Left Y-axis shows the normalized peak intensities of standard (blue) and GEN-HSQC (red) relative to the calibrated reference spectra (}{}$I_{RECT}^0$ and }{}$I_{GEN}^0$), whereas the right Y-axis represents the normalized intensity of standard HSQCs (blue) with respect to }{}$I_{GEN}^0$. (D) GEN-HSQC spectra of 200 µM 15N labeled catalytic subunit of protein kinase A (42 kDa). The figure insets compare the intensity of various resonances recorded using GEN-HSQC (red) and PFG-HSQC (blue) of the HSQC experiment. Both experiments are recorded on a Bruker 850 MHz spectrometer at 273 K using 128 complex indirect points were acquired with 16 scans per FID and a 1.5 s relaxation delay. Maximum RF amplitudes were 16.67 kHz for 1H and 7.7 kHz for 15N.CC BY-NC-ND · PubMed Central
Figure 5. Band-selective excitation and inversion pulses with (A) excitation on the left half of BW and (B) Inversion on left half of the BW. Note that to observe Mz and -Mz, we added a reading pulse (90o) prior to acquisition. (C) Excitation on left half and inversion on right half of the BW, simultaneously. The excitation/inversion half of the offset response can be swapped by inverting the phase array. An RF amplitude of 25 kHz was used for the simulation. Experimental verification was performed using a uniformly 13C labeled maltose binding protein (MBP) on an 850 MHz spectrometer. The band-selective excitation/inversion of aliphatic/carbonyl regions of the MBP are shown in the right column. The reference of the 13C spectra is shown in blue, whereas the spectra obtained with different GENETICS-AI band-selective operations are shown in red.
Figure 5. Band-selective excitation and inversion pulses with (A) excitation on the left half of BW and (B) Inversion on left half of the BW. Note that to observe Mz and -Mz, we added a reading pulse (90o) prior to acquisition. (C) Excitation on left half and inversion on right half of the BW, simultaneously. The excitation/inversion half of the offset response can be swapped by inverting the phase array. An RF amplitude of 25 kHz was used for the simulation. Experimental verification was performed using a uniformly 13C labeled maltose binding protein (MBP) on an 850 MHz spectrometer. The band-selective excitation/inversion of aliphatic/carbonyl regions of the MBP are shown in the right column. The reference of the 13C spectra is shown in blue, whereas the spectra obtained with different GENETICS-AI band-selective operations are shown in red.CC BY-NC-ND · PubMed Central
Figure 6. SE imaging under ultra-high RF inhomogeneity. (A) SE pulse sequence used for imaging. The π/2 and π pulses are replaced with GENETICS-AI pulses for GEN-SE sequence. The amplitude of the pulses is constant (16.67 kHz) and pulse lengths for π/2 and π pulses are 437 and 567 µs, respectively. The maximum RF amplitude tolerance level (}{}${\rm{\Delta \omega }}_1^{max}$) of these pulses were ±80%. The gradient strengths (G1 and G2) for all the experiments were set at 50% (3.3 G cm–1). (B) Diagram of phantom and coil geometry. The phantom consists of two glass beads of diameter 4 mm in a 5 mm Shigemi tube filled with 10% D2O and 90% H2O. (C) 1D image of the phantom using the pulse sequence in (A). (D) RF inhomogeneity response of 1D imaging using standard rectangular-shaped pulses of amplitude 16.67 kHz. (E) Same as (D) using GENETICS-AI pulses. The RF shapes and their fidelity profiles are shown in Figure S17. The experiments were performed in a Bruker 900 MHz NMR spectrometer.
Figure 6. SE imaging under ultra-high RF inhomogeneity. (A) SE pulse sequence used for imaging. The π/2 and π pulses are replaced with GENETICS-AI pulses for GEN-SE sequence. The amplitude of the pulses is constant (16.67 kHz) and pulse lengths for π/2 and π pulses are 437 and 567 µs, respectively. The maximum RF amplitude tolerance level (}{}${\rm{\Delta \omega }}_1^{max}$) of these pulses were ±80%. The gradient strengths (G1 and G2) for all the experiments were set at 50% (3.3 G cm–1). (B) Diagram of phantom and coil geometry. The phantom consists of two glass beads of diameter 4 mm in a 5 mm Shigemi tube filled with 10% D2O and 90% H2O. (C) 1D image of the phantom using the pulse sequence in (A). (D) RF inhomogeneity response of 1D imaging using standard rectangular-shaped pulses of amplitude 16.67 kHz. (E) Same as (D) using GENETICS-AI pulses. The RF shapes and their fidelity profiles are shown in Figure S17. The experiments were performed in a Bruker 900 MHz NMR spectrometer.CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

0 Valid · 1 Not valid · 23 Not checked · 85 Own results and interpretations

63 more claims

Create a free account to see all 109 claims and to verify the 52 references still open.

See all claims — free account
  • It may state two separate claims as one. 28
  • Probably not a substantial claim. 15

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

2 of 2

Citations

Cites

Show all 89 references

Research works printed but not linked: 1

  1. Cavanagh, 2007, Protein NMR spectroscopy: principles and practice, 1

Cited by

See all 24

Metadata

Authors
Subrahmanian, M. V., Pavuluri, K., Olivieri, C., & Veglia, G.
Affiliations
  • University of Minnesota (US)
  • Mayo Clinic (US)
  • Department of Biochemistry, Molecular Biology & Biophysics, University of Minnesota , Minneapolis, MN 55455, USA
Venue
PNAS Nexus
Published
2022-08-05
Volume
1
Issue
4
Pages
pgac133-pgac133
ISSN
2752-6542 (ISSN-L)
OpenAlex
W4289885698
Cited by
24
References
89
Citations by year
2026: 9 · 2025: 7 · 2024: 4 · 2023: 3 · 2022: 1
FWCI
1.18
Type
journal article
Language
en
Open access
Locations
  • PNAS Nexus (publishedVersion, cc-by-nc-nd)
  • Research Square (submittedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • Archivio Istituzionale della Ricerca (Universita Degli Studi Di Milano) (submittedVersion, other-oa)
Funders
National Institutes of Health (HL144130, GM64742, GM100310)
Topics
  • Physical Sciences › Chemistry › Spectroscopy › Advanced NMR Techniques and Applications
  • Physical Sciences › Computer Science › Artificial Intelligence › Quantum Computing Algorithms and Architecture
  • Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Advanced MRI Techniques and Applications
Keywords
spin entanglement, magnetic resonance imaging, quantum computing

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-18, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-18 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: High-fidelity control of spin ensemble dynamics via artificial intelligence: from quantum computing to NMR spectroscopy and imaging. https://app.cochat.ai/papers/10.1093/pnasnexus/pgac133
The paper
Subrahmanian, M. V., Pavuluri, K., Olivieri, C., & Veglia, G. (2022). High-fidelity control of spin ensemble dynamics via artificial intelligence: from quantum computing to NMR spectroscopy and imaging. PNAS Nexus. https://doi.org/10.1093/pnasnexus/pgac133

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.