Abstract

Biological neurons perform information processing using a model called pulse processing, which is both computationally efficient and scalable, adopting the best features of both analog and digital computing. Implementing pulse processing with photonics can result in bandwidths that are billions of times faster than biological neurons and substantially faster than electronics. Neurons have the ability to learn and adapt their processing based on experience through a change in the strength of synaptic connections in response to spiking activity. This mechanism is called spike-timing-dependent plasticity (STDP). Functionally, STDP constitutes a mechanism in which strengths of connections between neurons are based on the timing and order between presynaptic spikes and postsynaptic spikes, essentially forming a pulse lead/lag timing detector that is useful in feedback control and adaptation. Here we report for the first time the demonstration of optical STDP that is useful in pulse lead/lag timing detection and apply it to automatic gain control of a photonic pulse processor.

© 2013 Optical Society of America

Full Article  |  PDF Article

References

  • View by:
  • |
  • |
  • |

  1. J. J. Hopfield and D. W. Tank, Science 233, 625 (1986).
    [CrossRef]
  2. M. Mahowald and R. A. Douglas, Nature 354, 515 (1991).
    [CrossRef]
  3. D. Rosenbluth, K. Kravtsov, M. P. Fok, and P. R. Prucnal, Opt. Express 17, 22767 (2009).
    [CrossRef]
  4. M. P. Fok, D. Rosenbluth, K. Kravtsov, and P. R. Prucnal, IEEE Signal Process. Mag. 27, 160 (2010).
    [CrossRef]
  5. K. Kravtsov, M. P. Fok, D. Rosenbluth, and P. R. Prucnal, Opt. Express 19, 2133 (2011).
    [CrossRef]
  6. M. P. Fok, H. Deming, M. Nahmias, N. Rafidi, D. Rosenbluth, A. Tait, Y. Tian, and P. R. Prucnal, Opt. Lett. 36, 19 (2011).
    [CrossRef]
  7. S. Song, K. D. Miller, and L. F. Abbott, Nat. Neurosci. 3, 919 (2000).
    [CrossRef]
  8. R. C. Froemke and Y. Dan, Nature 416, 433 (2002).
    [CrossRef]
  9. N. Caporale and Y. Dan, Annu. Rev. Neurosci. 31, 25 (2008).
    [CrossRef]
  10. G. Bi and M. Poo, Annu. Rev. Neurosci. 24, 139 (2001).
    [CrossRef]
  11. W. Gerstner, R. Kempter, J. L. V. Hemmen, and H. Wagner, Nature 383, 76 (1996).
    [CrossRef]
  12. A. A. Minai and W. B. Levy, in INNS World Congress of Neural Networks II (Erlbaum, 1993), pp. 505–508.
  13. L. F. Abbott and K. I. Blum, Cereb. Cortex 6, 406 (1996).
    [CrossRef]
  14. P. D. Roberts, J. Comput. Neurosci. 7, 235 (1999).
    [CrossRef]
  15. K. I. Blum and L. F. Abbott, Neural Comput. 8, 85 (1996).
    [CrossRef]
  16. M. R. Mehta, M. C. Quirk, and M. Wilson, Neuron 25, 707 (2000).
    [CrossRef]
  17. R. Rao and T. J. Sejnowski, in Advances in Neural Information Processing Systems 12 (MIT, 2000), pp. 164–170.
  18. G. Chechik, Neurocomputation 15, 1481 (2003).
    [CrossRef]
  19. L. Buesing and W. Maass, in Advances in Neural Information Processing Systems 20 (MIT, 2007), pp. 193–200.
  20. C. Savin, P. Joshi, and J. Triesch, PLoS Comput. Biol. 6, e1000757 (2010).
    [CrossRef]
  21. M. Suzuki, H. Tanaka, and S. Akiba, Electron. Lett. 25, 88 (1989).
    [CrossRef]
  22. M. Koga, N. Tokura, and K. Nawata, Appl. Opt. 27, 3964 (1988).
    [CrossRef]

2011 (2)

2010 (2)

C. Savin, P. Joshi, and J. Triesch, PLoS Comput. Biol. 6, e1000757 (2010).
[CrossRef]

M. P. Fok, D. Rosenbluth, K. Kravtsov, and P. R. Prucnal, IEEE Signal Process. Mag. 27, 160 (2010).
[CrossRef]

2009 (1)

2008 (1)

N. Caporale and Y. Dan, Annu. Rev. Neurosci. 31, 25 (2008).
[CrossRef]

2003 (1)

G. Chechik, Neurocomputation 15, 1481 (2003).
[CrossRef]

2002 (1)

R. C. Froemke and Y. Dan, Nature 416, 433 (2002).
[CrossRef]

2001 (1)

G. Bi and M. Poo, Annu. Rev. Neurosci. 24, 139 (2001).
[CrossRef]

2000 (2)

M. R. Mehta, M. C. Quirk, and M. Wilson, Neuron 25, 707 (2000).
[CrossRef]

S. Song, K. D. Miller, and L. F. Abbott, Nat. Neurosci. 3, 919 (2000).
[CrossRef]

1999 (1)

P. D. Roberts, J. Comput. Neurosci. 7, 235 (1999).
[CrossRef]

1996 (3)

K. I. Blum and L. F. Abbott, Neural Comput. 8, 85 (1996).
[CrossRef]

W. Gerstner, R. Kempter, J. L. V. Hemmen, and H. Wagner, Nature 383, 76 (1996).
[CrossRef]

L. F. Abbott and K. I. Blum, Cereb. Cortex 6, 406 (1996).
[CrossRef]

1991 (1)

M. Mahowald and R. A. Douglas, Nature 354, 515 (1991).
[CrossRef]

1989 (1)

M. Suzuki, H. Tanaka, and S. Akiba, Electron. Lett. 25, 88 (1989).
[CrossRef]

1988 (1)

1986 (1)

J. J. Hopfield and D. W. Tank, Science 233, 625 (1986).
[CrossRef]

Abbott, L. F.

S. Song, K. D. Miller, and L. F. Abbott, Nat. Neurosci. 3, 919 (2000).
[CrossRef]

L. F. Abbott and K. I. Blum, Cereb. Cortex 6, 406 (1996).
[CrossRef]

K. I. Blum and L. F. Abbott, Neural Comput. 8, 85 (1996).
[CrossRef]

Akiba, S.

M. Suzuki, H. Tanaka, and S. Akiba, Electron. Lett. 25, 88 (1989).
[CrossRef]

Bi, G.

G. Bi and M. Poo, Annu. Rev. Neurosci. 24, 139 (2001).
[CrossRef]

Blum, K. I.

L. F. Abbott and K. I. Blum, Cereb. Cortex 6, 406 (1996).
[CrossRef]

K. I. Blum and L. F. Abbott, Neural Comput. 8, 85 (1996).
[CrossRef]

Buesing, L.

L. Buesing and W. Maass, in Advances in Neural Information Processing Systems 20 (MIT, 2007), pp. 193–200.

Caporale, N.

N. Caporale and Y. Dan, Annu. Rev. Neurosci. 31, 25 (2008).
[CrossRef]

Chechik, G.

G. Chechik, Neurocomputation 15, 1481 (2003).
[CrossRef]

Dan, Y.

N. Caporale and Y. Dan, Annu. Rev. Neurosci. 31, 25 (2008).
[CrossRef]

R. C. Froemke and Y. Dan, Nature 416, 433 (2002).
[CrossRef]

Deming, H.

Douglas, R. A.

M. Mahowald and R. A. Douglas, Nature 354, 515 (1991).
[CrossRef]

Fok, M. P.

Froemke, R. C.

R. C. Froemke and Y. Dan, Nature 416, 433 (2002).
[CrossRef]

Gerstner, W.

W. Gerstner, R. Kempter, J. L. V. Hemmen, and H. Wagner, Nature 383, 76 (1996).
[CrossRef]

Hemmen, J. L. V.

W. Gerstner, R. Kempter, J. L. V. Hemmen, and H. Wagner, Nature 383, 76 (1996).
[CrossRef]

Hopfield, J. J.

J. J. Hopfield and D. W. Tank, Science 233, 625 (1986).
[CrossRef]

Joshi, P.

C. Savin, P. Joshi, and J. Triesch, PLoS Comput. Biol. 6, e1000757 (2010).
[CrossRef]

Kempter, R.

W. Gerstner, R. Kempter, J. L. V. Hemmen, and H. Wagner, Nature 383, 76 (1996).
[CrossRef]

Koga, M.

Kravtsov, K.

Levy, W. B.

A. A. Minai and W. B. Levy, in INNS World Congress of Neural Networks II (Erlbaum, 1993), pp. 505–508.

Maass, W.

L. Buesing and W. Maass, in Advances in Neural Information Processing Systems 20 (MIT, 2007), pp. 193–200.

Mahowald, M.

M. Mahowald and R. A. Douglas, Nature 354, 515 (1991).
[CrossRef]

Mehta, M. R.

M. R. Mehta, M. C. Quirk, and M. Wilson, Neuron 25, 707 (2000).
[CrossRef]

Miller, K. D.

S. Song, K. D. Miller, and L. F. Abbott, Nat. Neurosci. 3, 919 (2000).
[CrossRef]

Minai, A. A.

A. A. Minai and W. B. Levy, in INNS World Congress of Neural Networks II (Erlbaum, 1993), pp. 505–508.

Nahmias, M.

Nawata, K.

Poo, M.

G. Bi and M. Poo, Annu. Rev. Neurosci. 24, 139 (2001).
[CrossRef]

Prucnal, P. R.

Quirk, M. C.

M. R. Mehta, M. C. Quirk, and M. Wilson, Neuron 25, 707 (2000).
[CrossRef]

Rafidi, N.

Rao, R.

R. Rao and T. J. Sejnowski, in Advances in Neural Information Processing Systems 12 (MIT, 2000), pp. 164–170.

Roberts, P. D.

P. D. Roberts, J. Comput. Neurosci. 7, 235 (1999).
[CrossRef]

Rosenbluth, D.

Savin, C.

C. Savin, P. Joshi, and J. Triesch, PLoS Comput. Biol. 6, e1000757 (2010).
[CrossRef]

Sejnowski, T. J.

R. Rao and T. J. Sejnowski, in Advances in Neural Information Processing Systems 12 (MIT, 2000), pp. 164–170.

Song, S.

S. Song, K. D. Miller, and L. F. Abbott, Nat. Neurosci. 3, 919 (2000).
[CrossRef]

Suzuki, M.

M. Suzuki, H. Tanaka, and S. Akiba, Electron. Lett. 25, 88 (1989).
[CrossRef]

Tait, A.

Tanaka, H.

M. Suzuki, H. Tanaka, and S. Akiba, Electron. Lett. 25, 88 (1989).
[CrossRef]

Tank, D. W.

J. J. Hopfield and D. W. Tank, Science 233, 625 (1986).
[CrossRef]

Tian, Y.

Tokura, N.

Triesch, J.

C. Savin, P. Joshi, and J. Triesch, PLoS Comput. Biol. 6, e1000757 (2010).
[CrossRef]

Wagner, H.

W. Gerstner, R. Kempter, J. L. V. Hemmen, and H. Wagner, Nature 383, 76 (1996).
[CrossRef]

Wilson, M.

M. R. Mehta, M. C. Quirk, and M. Wilson, Neuron 25, 707 (2000).
[CrossRef]

Annu. Rev. Neurosci. (2)

N. Caporale and Y. Dan, Annu. Rev. Neurosci. 31, 25 (2008).
[CrossRef]

G. Bi and M. Poo, Annu. Rev. Neurosci. 24, 139 (2001).
[CrossRef]

Appl. Opt. (1)

Cereb. Cortex (1)

L. F. Abbott and K. I. Blum, Cereb. Cortex 6, 406 (1996).
[CrossRef]

Electron. Lett. (1)

M. Suzuki, H. Tanaka, and S. Akiba, Electron. Lett. 25, 88 (1989).
[CrossRef]

IEEE Signal Process. Mag. (1)

M. P. Fok, D. Rosenbluth, K. Kravtsov, and P. R. Prucnal, IEEE Signal Process. Mag. 27, 160 (2010).
[CrossRef]

J. Comput. Neurosci. (1)

P. D. Roberts, J. Comput. Neurosci. 7, 235 (1999).
[CrossRef]

Nat. Neurosci. (1)

S. Song, K. D. Miller, and L. F. Abbott, Nat. Neurosci. 3, 919 (2000).
[CrossRef]

Nature (3)

R. C. Froemke and Y. Dan, Nature 416, 433 (2002).
[CrossRef]

M. Mahowald and R. A. Douglas, Nature 354, 515 (1991).
[CrossRef]

W. Gerstner, R. Kempter, J. L. V. Hemmen, and H. Wagner, Nature 383, 76 (1996).
[CrossRef]

Neural Comput. (1)

K. I. Blum and L. F. Abbott, Neural Comput. 8, 85 (1996).
[CrossRef]

Neurocomputation (1)

G. Chechik, Neurocomputation 15, 1481 (2003).
[CrossRef]

Neuron (1)

M. R. Mehta, M. C. Quirk, and M. Wilson, Neuron 25, 707 (2000).
[CrossRef]

Opt. Express (2)

Opt. Lett. (1)

PLoS Comput. Biol. (1)

C. Savin, P. Joshi, and J. Triesch, PLoS Comput. Biol. 6, e1000757 (2010).
[CrossRef]

Science (1)

J. J. Hopfield and D. W. Tank, Science 233, 625 (1986).
[CrossRef]

Other (3)

A. A. Minai and W. B. Levy, in INNS World Congress of Neural Networks II (Erlbaum, 1993), pp. 505–508.

R. Rao and T. J. Sejnowski, in Advances in Neural Information Processing Systems 12 (MIT, 2000), pp. 164–170.

L. Buesing and W. Maass, in Advances in Neural Information Processing Systems 20 (MIT, 2007), pp. 193–200.

Cited By

OSA participates in CrossRef's Cited-By Linking service. Citing articles from OSA journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (5)

Fig. 1.
Fig. 1.

(a) Unsupervised learning with STDP. (b) STDP characteristic. tposttpre: time interval between the post- and presynaptic spikes. ΔW: resultant change in synaptic strength.

Fig. 2.
Fig. 2.

(a) Optical implementation of STDP. (b) Formation of depression window. (c) Formation of potentiation window. (d) Linearly combining (b) and (c) results in STDP characteristic.

Fig. 3.
Fig. 3.

Experimentally measured optical STDP characteristic.

Fig. 4.
Fig. 4.

(a) Reconfigurable STDP characteristic. (b) Different SOA driving current. (c) Different pre- and postsynaptic spikes splitting ratio. (d) Different input power to the EAM.

Fig. 5.
Fig. 5.

Experimental results of automatic gain control in pulse-processing device based on optical STDP. (a) Correct response presented by the teacher; (b)–(d) initial weight is too low, and the optical STDP adjusts its gain automatically; (e)–(g) initial weight is too high, and the optical STDP adjust its gain automatically.

Metrics