Jan Antolik PhD

Jan Antolik


Research Interests

I have worked on biologically realistic computational models of early visual system. Important aspects of my work is that these models are strongly constrained by our knowledge about structural connectivity in the visual system and also that these models capture the development of the cortex, thus are able to explain how these complicated structures can arise from initial random conditions through few simple rules. My main motivation was to understand the biological system - the brain, however from a long term perspective I intend to go back to my original background - Artificial Intelligence - by transferring the gained knowledge into building artificial systems.

Publications:
2013
  Developing orientation maps using realistic patterns of lateral connectivity
Rudiger, P, Law, J, Antolik, J & Bednar, J 2013, 'Developing orientation maps using realistic patterns of lateral connectivity' 22nd Annual Computational Neuroscience Meeting: CNS 2013, Paris, France, 13/07/13 - 18/07/13, .
While developmental models have been very successful in replicating the main features of experimentally observed topographic maps in the primary visual cortex (V1), they have relied on several unrealistic assumptions. These models are typically variants of the self-organizing
map model [1], and almost universally assume “Mexicanhat” lateral connectivity in V1, with short-range excitatory and longer-range inhibitory connections. Experimental data is in direct conflict with this assumption, with anatomical tracing studies showing that neurons making long-range connections are excitatory [2,3]. A variety of
electrophysiological and psychophysical studies also suggest both excitatory and inhibitory effects at long ranges, depending on experimental conditions. The current consensus is that the actual pattern of connectivity consists of long-range excitation leading to di-synaptic inhibition via local inhibitory interneurons [2-4]. The
resulting aggregate circuit has an overall inhibitory effect when the excitatory drive to local inhibitory synapses is large enough. In principle, the behavior of this circuit at high input contrasts may therefore mimic the Mexican-hat profile of these earlier model, while potentially exhibiting more realistic contrast dependent behavior.

We present a rate-based model of simple-cell development that robustly self-organizes into biologically realistic orientation maps on the basis of this experimentally determined connectivity. The model is built using the Topographica simulator [5], and consists of a number of sheets of units representing the retinal photoreceptors, RGC/LGN
cells, and individual populations of excitatory and inhibitory V1 neurons. The receptive field weights, initialized randomly within a Gaussian envelope, are adjusted through Hebbian learning with divisive normalization in response to activity driven by 20,000 consecutive input
patterns (either natural images or artificial patterns). We show that development of realistic maps is robust, primarily due to homeostaticmechanisms in V1 and divisive contrast-gain control in the RGC/LGNlayer.

The model demonstrates that the experimentally established connectivity framework can lead to orderly map development and can replicate many of the contextual and contrast dependent effects observed in adult V1. This work looks at how Mexican-hat connectivity arises from the overall network interactions at high contrast and how
it adjusts at lower contrasts. Further, it demonstrates clearly how patchy long-range connectivity between isoorientation domains emerges, and the role it plays in modulating V1 activity. In doing so, the model provides a clear link between topographic map formation, the development of the underlying connectivity, and the perceptual consequences of this circuitry, including contrast-dependent
size-tuning shifts and the early stages of more complex effects like pop-out and contour completion.

In future, this work will help us to complete our understanding of the V1 circuit by adding feedback mechanisms or selectively modulating specific connections to model the effects of different neuromodulators. Additionally, the results may be used to provide realistic connectivity patterns for large scale spiking models, which often struggle to adequately constrain their connectivity. Overall, this model demonstrates for the first time that it is possible to robustly develop biologically plausible orientation maps on the basis of realistic connectivity, accounting for various surround modulation effects and providing a solid basis for future models of V1.

References

[1] Von der Malsburg, C.: Self-organization of orientation sensitive cells in the striate cortex. Kybernetik 1973, 14(2): 85–100

[2] Gilbert, D., & Wiesel, T.: Columnar specificity of intrinsic horizontal and corticocortical connections in cat visual cortex. The Journal of Neuroscience 1989, 9(7): 2432–2442.

[3] Hirsch, J. A, & Gilbert, C. D.: Synaptic physiology of horizontal connections in the cat’s visual cortex. The Journal of Neuroscience 1991, 11(6): 1800–9

[4] Weliky, M., Kandler, K., Fitzpatrick, D., & Katz, L. C. : Patterns of excitation and inhibition evoked by horizontal connections in visual cortex share a common relationship to orientation columns. Neuron 1995, 15(3): 541–52

[5] The Topographica Neural Map Simulator. [http://www.topographica.org]
General Information
Organisations: Institute for Adaptive and Neural Computation .
Authors: Rudiger, Philipp, Law, Judith, Antolik, Jan & Bednar, James.
Publication Date: 2013
Publication Information
Category: Poster
Original Language: English
  Mechanisms for Stable, Robust, and Adaptive Development of Orientation Maps in the Primary Visual Cortex
Stevens, J-L, Law, J, Antolik, J & Bednar, JA 2013, 'Mechanisms for Stable, Robust, and Adaptive Development of Orientation Maps in the Primary Visual Cortex' Journal of Neuroscience, vol 33, no. 40, pp. 15747-15766. DOI: 10.1523/JNEUROSCI.1037-13.2013
Development of orientation maps in ferret and cat primary visual cortex (V1) has been shown to be stable, in that the earliest measurable maps are similar in form to the eventual adult map, robust, in that similar maps develop in both dark rearing and in a variety of normal visual environments, and yet adaptive, in that the final map pattern reflects the statistics of the specific visual environment. How can these three properties be reconciled? Using mechanistic models of the development of neural connectivity in V1, we show for the first time that realistic stable, robust, and adaptive map development can be achieved by including two low-level mechanisms originally motivated from single-neuron results. Specifically, contrast-gain control in the retinal ganglion cells and the lateral geniculate nucleus reduces variation in the presynaptic drive due to differences in input patterns, while homeostatic plasticity of V1 neuron excitability reduces the postsynaptic variability in firing rates. Together these two mechanisms, thought to be applicable across sensory systems in general, lead to biological maps that develop stably and robustly, yet adapt to the visual environment. The modeling results suggest that topographic map stability is a natural outcome of low-level processes of adaptation and normalization. The resulting model is more realistic, simpler, and far more robust, and is thus a good starting point for future studies of cortical map development.
General Information
Organisations: Institute for Adaptive and Neural Computation .
Authors: Stevens, Jean-Luc, Law, Judith, Antolik, Jan & Bednar, James A. .
Number of pages: 20
Pages: 15747-15766
Publication Date: 2 Oct 2013
Publication Information
Category: Article
Journal: Journal of Neuroscience
Volume: 33
Issue number: 40
ISSN: 0270-6474
Original Language: English
DOIs: 10.1523/JNEUROSCI.1037-13.2013
2011
  Stable and robust development of orientation maps and receptive fields
Law, JS, Antolik, J & Bednar, JA 2011, 'Stable and robust development of orientation maps and receptive fields' BMC Neuroscience, vol 12, no. Suppl 1, pp. P10. DOI: 10.1186/1471-2202-12-S1-P10
General Information
Organisations: Neuroinformatics DTC.
Authors: Law, Judith S, Antolik, Jan & Bednar, James A.
Pages: P10
Publication Date: 1 Jan 2011
Publication Information
Category: Article
Journal: BMC Neuroscience
Volume: 12
Issue number: Suppl 1
ISSN: 1471-2202
Original Language: English
DOIs: 10.1186/1471-2202-12-S1-P10
  Development of maps of simple and complex cells in the primary visual cortex
Antolik, J & Bednar, JA 2011, 'Development of maps of simple and complex cells in the primary visual cortex' Frontiers in Computational Neuroscience, vol 5, no. 17. DOI: 10.3389/fncom.2011.00017
Hubel & Wiesel classified primary visual cortex (V1) neurons as either simple, with responses modulated by the spatial phase of a sine grating, or complex, i.e. largely phase invariant. Much progress has been made in understanding how simple cells develop, and there are now detailed computational models establishing how they can form topographic maps ordered by orientation preference. There are also models of how complex cells can develop using outputs from simple cells with different phase preferences, but no model of how a topographic orientation map of complex cells could be formed based on the actual connectivity patterns found in V1. Addressing this question is important, because the majority of existing developmental models of simple-cell maps group neurons selective to similar spatial phases together, which is contrary to experimental evidence, and makes it difficult to construct complex cells. Overcoming this limitation is not trivial, because mechanisms responsible for map development drive receptive fields of nearby neurons to be highly correlated, while co-oriented receptive fields of opposite phases are anti-correlated. In this work, we model V1 as two topographically organised sheets representing cortical layer 4 and 2/3. Only layer 4 receives direct thalamic input. Both sheets are connected with narrow feed-forward and feedback connectivity. Only layer 2/3 contains strong long-range lateral connectivity, in line with current anatomical findings. Initially all weights in the model are random, and each is modified via a Hebbian learning rule. The model develops smooth, matching, orientation preference maps in both sheets. Layer 4 units become simple cells, with phase preference arranged randomly, while those in layer 2/3 are primarily complex cells. To our knowledge this model is the first explaining how simple cells can develop with random phase preference, and how maps of complex cells can develop, using only realistic patterns of connectivity.
General Information
Organisations: Institute for Adaptive and Neural Computation .
Authors: Antolik, Jan & Bednar, James A.
Keywords: (, , . )
Publication Date: 2011
Publication Information
Category: Article
Journal: Frontiers in Computational Neuroscience
Volume: 5
Issue number: 17
ISSN: 1662-5188
Original Language: English
DOIs: 10.3389/fncom.2011.00017
2009
  Reconciling models of V1 development and adult function
Antolik, J, Law, JS & Bednar, J 2009, 'Reconciling models of V1 development and adult function' Society for Neuroscience Annual Meeting 2009, Chicago, United States, 17/10/09 - 21/10/09, .
General Information
Organisations: Institute for Adaptive and Neural Computation .
Authors: Antolik, Jan, Law, Judith S. & Bednar, James.
Publication Date: 2009
Publication Information
Category: Poster
Original Language: English
  Homeostatic and gain control mechanisms in a developmental model of orientation map formation in V1
Law, J, Bednar, J & Antolik, J 2009, 'Homeostatic and gain control mechanisms in a developmental model of orientation map formation in V1' Computational and Systems Neuroscience (Cosyne) 2009, Salt Lake City, UT, United States, 26/02/09 - 3/03/09, . DOI: 10.3389/conf.neuro.06.2009.03.341
Numerous studies have shown that cortical neurons can self-regulate their response gain (i.e., their output in response to an input). Theoretical studies of such gain control have primarily considered single cells or small networks of neurons in the adult brain. However, gain control is likely to be particularly important during development, because the amount and distribution of input activity can change dramatically between neurogenesis and adulthood. For instance, the developing visual system at first receives intrinsically generated input, such as retinal waves or spontaneous cortical activity, and in later stages (after eye opening) receives direct visual stimulation from the environment. In this study we examine how gain control can interact with basic homeostatic mechanisms to reproduce the experimentally observed patterns of development in a large scale model of an orientation map in the primary visual cortex (V1). Using this model, we have identified a small set of mathematical rules that can reproduce the following experimentally observed phenomena: stable orientation map development (Chapman et al. J. Neurosci., 1996, 16:6443--6453), contrast independent orientation tuning (Alitto et al. J Neurophysiol., 2004 91:2797--2808), and orientation map development that is robust against changes in the levels or distributions of input activity over time (Crair et al. Science, 1998, 279:566--570). We show that the above constraints can be met by using a simple but plausible gain control mechanism at the level of the Lateral Geniculate Nucleus (LGN) or retina, plus a mechanism that maintains a constant ratio between the strength of different input types (afferent vs. feedback, and excitatory vs. inhibitory) to each individual neuron. By directly maintaining these specific interaction ratios, it is sufficient to use a simple threshold adjustment rule for each neuron, rather than the more complex intrinsic excitability adjustment rules previously designed for more abstract networks (Triesch ICANN, 2005, 65--70). This model thus highlights the benefit of studying these phenomena in neural models whose architecture is constrained by the known connectivity of neural structures (such as V1).
General Information
Organisations: Institute for Adaptive and Neural Computation .
Authors: Law, Judith, Bednar, James & Antolik, Jan.
Publication Date: 2009
Publication Information
Category: Poster
Original Language: English
DOIs: 10.3389/conf.neuro.06.2009.03.341
2008
  Developing maps of complex cells in a computational model of V1
Antolik, J & Bednar, J 2008, 'Developing maps of complex cells in a computational model of V1' Society for Neuroscience Annual meeting, 2008, Washington DC, United States, 15/11/08 - 19/11/08, .
General Information
Organisations: Institute for Adaptive and Neural Computation .
Authors: Antolik, Jan & Bednar, James.
Publication Date: 19 Nov 2008
Publication Information
Category: Poster
Original Language: English
  Developing maps of complex cells in a computational model
Antolik, J & Bednar, J 2008, 'Developing maps of complex cells in a computational model' 6th FENS Forum of European Neuroscience, Geneva, Switzerland, 12/07/08 - 16/07/08, .
Hubel & Wiesel (1968) classified primary visual cortex (V1) neurons as either simple, with responses strongly modulated by the spatial phase of a sine grating, or complex, i.e. largely phase invariant. Much progress has been made in understanding how simple cells develop, and there are now detailed computational models establishing how they can form topographic maps ordered by orientation preference. There are also models of how individual complex cells can develop using outputs from simple cells with different phase preferences, but no model of how a realistic topographic orientation map of complex cells could be formed based on the actual connectivity patterns found in V1. Addressing this question is important, because existing simple-cell models produce maps that group similar spatial phases together, which is contrary to experimental evidence, and makes it difficult to construct complex cells. Overcoming this limitation is not trivial, because the simple-cell models are driven by correlations in the input, and phase is more highly correlated than orientation in natural images.
In this work, we model V1 as two topographically organised sheets, one representing cortical layer 4c and one representing layer 2/3. Only layer 4c receives direct thalamic input. Both sheets are connected with narrow feed-forward and feedback connectivity. Only layer 2/3 contains long range lateral connectivity, in line with current anatomical findings. Initially all weights in the model are random, and each is modified via a Hebbian learning rule. The model develops smooth, matching, realistic orientation preference maps in both sheets. Layer 4c units become simple cells, with phase preference arranged randomly, while those in layer 2/3 are primarily complex cells. To our knowledge this model is the first explaining how simple cells can develop with random phase preference, and how smoothly organised maps of complex cells can develop, using only realistic patterns of connectivity.
General Information
Organisations: Institute for Adaptive and Neural Computation .
Authors: Antolik, Jan & Bednar, James.
Publication Date: 13 Jul 2008
Publication Information
Category: Poster
Original Language: English
2007
  Modelling surround modulation in the LGN
Antolik, J & Bednar, J 2007, 'Modelling surround modulation in the LGN' Society for Neuroscience Annual Meeting, 2007, San Diego, California, United States, 3/11/07 - 7/11/07, .
General Information
Organisations: Neuroinformatics DTC.
Authors: Antolik, Jan & Bednar, James.
Publication Date: Nov 2007
Publication Information
Category: Poster
Original Language: English
  Modeling the development of maps of complex cells in V1
Antolik, J & Bednar, J 2007, 'Modeling the development of maps of complex cells in V1' Neurons on the Map - Fifth Annual Neuron Satellite Meeting, San Diego, CA, United States, 1/11/07, .
General Information
Organisations: Neuroinformatics DTC.
Authors: Antolik, Jan & Bednar, James.
Publication Date: 2007
Publication Information
Category: Abstract
Original Language: English

Projects:
Modelling the role of extra-striate feedback in contextual modulation of V1 (PhD)