Quantum information is fragile: the slightest noise can destroy it. That's why two protection methods have been devised. The first is dynamical decoupling: rapid pulses are applied to qubits to average the noise to zero. It's like shaking a box of fragile items so they don't bump against the walls. The second is quantum error correction: information is encoded in multiple qubits, and then special checks are measured to detect errors without destroying the data itself.

But you can't simply combine these methods: dynamical decoupling can disrupt the code space, and error correction works with the errors that remain after decoupling. The authors proposed doing decoupling logically: that is, using operations that preserve the code space. They derived formulas showing when such a hybrid outperforms each method individually.

For the Steane code [[7,1,3]] and the code [[13,1,3]], they numerically found parameter regions where the hybrid is better. This shows that the code, decoder, and decoupling group should be designed jointly, not separately.