Neural Network Circuit for Operant Conditioning with Blocking and Overshadowing Effects Based on DNA Strand Displacement
Operant conditioning reflects the ability of organisms to adapt and learn. By implementing operant conditioning neural networks, complex brain-like behaviors can be simulated and learned at the molecular level. In this paper, an operant conditioning circuit with blocking effect and overshadowing effect is proposed by DNA strand displacement technique. The network can achieve blocking and overshadowing effects in the presence of multiple inputs. First, we construct DNA memristors using DNA strand displacement technology. The constructed DNA memristors are plastic and nonvolatile, which is very similar to biological synapses. The synaptic circuits constructed by DNA memristors have short-term plasticity, long-term plasticity, potentiation plasticity and depression plasticity. Second, the learning module, forgetting module, synaptic module, decision module and feedback module are constructed, and the operant conditioning circuit is realized. In addition, the operant conditioning circuit has relearning effect. Blocking and masking effects can also be achieved in the presence of multiple inputs. Finally, the reliability of the circuit is verified by the simulation of Visual DSD software. Our work demonstrates the potential of DNA molecules to build complex intelligent systems and provides an idea for using DNA molecules to achieve artificial intelligence.
Authors
- Junwei Sun (ORCID: https://orcid.org/0000-0001-8518-5064)
- Jinjiang Wang (ORCID: https://orcid.org/0000-0003-0163-4446)
- Yan Wang (ORCID: https://orcid.org/0000-0003-4224-5391)
- Jiaming Li (ORCID: https://orcid.org/0000-0003-3558-3560)
- Yanfeng Wang (ORCID: https://orcid.org/0000-0001-7873-6577)
Institutions
- Zhengzhou University of Light Industry (CN)
Publication Details
- Journal
- ACS Synthetic Biology
- Published
- 2025-12-05
- DOI
- https://doi.org/10.1021/acssynbio.5c00672
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00