Computational Maps in the Visual Cortex
About the Authors
Back Cover    
Table of Contents 
Sample Chapter 
Purchase online at:
Click on the figure title to see the figure and the caption; click further on the image to see a PDF version of the figure. The PDF version can be zoomed in to see detail not visible in print.

Chapter 1   Introduction
1.1     Columnar organization of the primary visual cortex 
1.2     Spontaneous activity in the retina 
1.3     Perceptual grouping tasks 
1.4     Basic LISSOM model of the primary visual cortex 

Chapter 2   Biological Background
2.1     Human visual pathways (top view) 
2.2     Receptive field types in retina, LGN and V1 
2.3     Measuring cortical maps 
2.4     Orientation map in the macaque 
2.5     Hierarchical organization of feature preferences in the macaque 
2.6     Long-range lateral connections in the macaque 
2.7     Lateral connections in the tree shrew orientation map 
2.8     Spontaneous activity in the cat PGO pathway 
2.9     Solving the superposition catastrophe through temporal coding 
2.10     Synchronization of one and two input objects in the cat 

Chapter 3   Computational Foundations
3.1     Computational abstractions of neurons and networks 
3.2     Perceptual grouping through temporal coding 
3.3     General architecture of self-organizing map models of the primary visual cortex 
3.4     Training a self-organizing map with Gaussian activity patterns 
3.5     Self-organization of weight vectors 
3.6     Self-organization of a retinotopic map 
3.7     Magnification of dense input areas 
3.8     Principal components of data distributions 
3.9     Approximating nonlinear distributions with principal curves and folding 
3.10     Three-dimensional model of ocular dominance 

Chapter 4   LISSOM: A Computational Map Model of V1
4.1     Architecture of the basic LISSOM model 
4.2     Afferent weights of ON and OFF neurons in the LGN 
4.3     Initial V1 afferent and lateral weights 
4.4     Example input and response 
4.5     Neuron activation function σ (s) 
4.6     Self-organized V1 afferent weights 
4.7     Self-organized afferent and lateral weights across V1 
4.8     Self-organization of the retinotopic map 
4.9     Self-organized V1 lateral weights 

Chapter 5   Development of Maps and Connections
5.1     Fourier spectrum and gradient of the macaque orientation map 
5.2     Normal vs. strabismic cat ocular dominance maps and lateral connections 
5.3     Combined OR/OD map in the macaque 
5.4     Spatiotemporal receptive fields, direction maps, and combined OR/DR maps in animals 
5.5     Initial V1 afferent and lateral weights 
5.6     Example input and response 
5.7     Self-organized V1 afferent and lateral weights 
5.8     Self-organized afferent and lateral weights across V1 
5.9     Self-organization of the orientation map 
5.10     Fourier spectrum and gradient of the orientation map 
5.11     Retinotopic organization of the orientation map 
5.12     Long-range lateral connections in the orientation map 
5.13     Effect of training patterns on orientation maps 
5.14     LISSOM model of ocular dominance 
5.15     Self-organization of afferent weights into OD receptive fields 
5.16     Self-organized ocular dominance map 
5.17     Long-range lateral connections in the ocular dominance map 
5.18     Ocular dominance and long-range lateral connections in the strabismic ocular dominance map 
5.19     Effect of disparity on ocular dominance maps 
5.20     LISSOM model of orientation and direction selectivity 
5.21     Self-organization of afferent weights into spatiotemporal RFs 
5.22     Self-organized OR/DR map 
5.23     Combined OR/DR map 
5.24     Long-range lateral connections in the combined OR/DR map 
5.25     Effect of input speed on direction maps 
5.26     LISSOM model of orientation, ocular dominance, and direction selectivity 
5.27     Self-organized OR/OD map 
5.28     Long-range lateral connections in the combined OR/OD map 
5.29     Combined OR/OD/DR map trained with Gaussians 
5.30     Example natural image input for training the OR/OD/DR map 
5.31     Combined OR/OD/DR map trained with natural images 
5.32     Effect of training patterns on OR/OD/DR maps 

Chapter 6   Understanding Plasticity
6.1     Reorganization of receptive fields after a retinal lesion 
6.2     Reorganization of receptive fields after a cortical lesion 
6.3     Architecture of the reduced LISSOM model 
6.4     Effect of ON/OFF channels on orientation maps 
6.5     Role of ON/OFF channels in processing various kinds of inputs 
6.6     Retinal activation and V1 response before and after a retinal scotoma 
6.7     Reorganization of the orientation map after a retinal scotoma 
6.8     Dynamic RF expansion and perceptual shift after a retinal scotoma 
6.9     Retinal activation and V1 response before and after a cortical lesion 
6.10     Cortical response after a cortical lesion 
6.11     Reorganization of lateral inhibitory weights after a cortical lesion 
6.12     Reorganization of the orientation map after a cortical lesion 

Chapter 7   Understanding Visual Performance: The Tilt Aftereffect
7.1     Demonstration of the tilt aftereffect 
7.2     Tilt aftereffect in human subjects 
7.3     Measuring perceived orientation as vector sum 
7.4     Cortical response and perceived orientation 
7.5     Tilt aftereffect in humans and in LISSOM 
7.6     Tilt aftereffect over time in humans and in LISSOM 
7.7     Components of the tilt aftereffect due to each weight type 
7.8     Changes in lateral inhibitory weights due to adaptation 
7.9     Cortical response during adaptation and during direct and indirect tilt aftereffect 

Chapter 8   HLISSOM: A Hierarchical Model
8.1     Architecture of the HLISSOM model 
8.2     Effect of afferent normalization on V1 responses 
8.3     Effect of afferent normalization on V1 neuron tuning 
8.4     Internally generated and environmental input patterns 
8.5     Effect of different input streams and initial organizations on the self-organizing process 

Chapter 9   Understanding Low-Level Development: Orientation Maps
9.1     Effect of internally generated prenatal training patterns on orientation maps 
9.2     Prenatal orientation maps in animals and in HLISSOM 
9.3     Effect of environmental postnatal training patterns on orientation maps 
9.4     Postnatal orientation maps in animals and in HLISSOM 
9.5     Distribution of orientation preferences in animals and in HLISSOM 
9.6     Effect of prenatal and postnatal training on orientation maps 

Chapter 10   Understanding High-Level Development: Face Detection
10.1     Measuring newborn face preferences 
10.2     Face preferences in newborns 
10.3     Face preferences in young infants 
10.4     Self-organization of the scaled-up orientation map 
10.5     Self-organization of the FSA map 
10.6     Response to schematic images by Goren et al. (1975) and Johnson et al. (1991) 
10.7     Response to schematic images by Valenza et al. (1996) and Simion et al. (1998a) 
10.8     Spurious responses to the inverted three-dot pattern 
10.9     Response to natural images 
10.10     Response variation with size and viewpoint 
10.11     Effect of training patterns on face preferences 
10.12     Initial afferent weights across prenatally trained and naive FSA networks 
10.13     Face and object images in postnatal training 
10.14     Example postnatal training presentations 
10.15     Prenatally established bias for learning faces 
10.16     Postnatal decline in response to schematic images 
10.17     Mother preferences based on both internal and external features 

Chapter 11   PGLISSOM: A Perceptual Grouping Model
11.1     Architecture of the PGLISSOM model 
11.2     The leaky integrator neuron model 
11.3     Self-organized afferent weights and retinotopic organization 
11.4     Self-organized orientation map 
11.5     Long-range lateral connections in GMAP 
11.6     Activating neurons with collinear and cocircular RFs 
11.7     Distribution of lateral connections in animals and in PGLISSOM 

Chapter 12   Temporal Coding
12.1     Synchronized and desynchronized modes of firing 
12.2     Effect of connection type and decay rate on synchronization 
12.3     Effect of excitatory connection range on synchronization 
12.4     Binding and segmentation with different connection types 
12.5     Effect of noise on desynchronization 
12.6     Effect of relative input size on synchronization 
12.7     Overcoming noise with strong excitation 
12.8     Overcoming noise with a long refractory period 

Chapter 13   Understanding Perceptual Grouping: Contour Integration
13.1     Demonstration of contour integration 
13.2     Association fields for contour integration 
13.3     Edge-induced vs. line-end-induced illusory contours 
13.4     Contour completion across edge inducers 
13.5     Measuring local response as multi-unit activity 
13.6     Contour integration process with varying degrees of orientation jitter 
13.7     Contour integration performance in humans and in PGLISSOM 
13.8     Quantifying the spatial relationship between two receptive fields 
13.9     Edge cooccurrence in nature and long-range lateral connections in PGLISSOM 
13.10     Contour segmentation process 
13.11     Contour segmentation performance 
13.12     Contour completion process 
13.13     Afferent contribution in contour completion 
13.14     Contour completion process with different kinds of connections 
13.15     Contour completion performance with different kinds of connections 
13.16     Contour completion process in the illusory triangle 
13.17     Salience of complete vs. incomplete illusory triangles 
13.18     Contour completion performance in the illusory triangle 
13.19     Contour completion performance in closed vs. open contours 
13.20     Orientation selectivity in SMAP with different input distributions 
13.21     Lateral excitatory connections in GMAP with different input frequencies 
13.22     Lateral excitatory connections in GMAP with different curvature ranges 
13.23     Contour integration process with different input frequencies 
13.24     Contour integration process with different curvature ranges 
13.25     Contour integration performance with different input distributions 

Chapter 14   Computations in Visual Maps
14.1     Self-organized vs. isotropic lateral connections 
14.2     Sparse, redundancy-reduced coding with self-organized lateral connections 
14.3     Architecture of the handwritten digit recognition system 
14.4     Handwritten digit examples 
14.5     Self-organized SOM afferent weights 
14.6     Self-organized LISSOM afferent and lateral weights 
14.7     SOM activity patterns 
14.8     LISSOM activity patterns 

Chapter 15   Scaling LISSOM simulations
15.1     Scaling retinal and cortical area 
15.2     Scaling retinal density 
15.3     Scaling cortical density 
15.4     Training time and memory usage in LISSOM vs. GLISSOM 
15.5     Weight interpolation in GLISSOM 
15.6     Scaling cortical density in GLISSOM 
15.7     Self-organization of LISSOM and GLISSOM orientation maps 
15.8     Accuracy of the final GLISSOM map as a function of the initial network size 
15.9     Orientation maps in LISSOM and GLISSOM 
15.10     Simulation time and memory usage in LISSOM vs. GLISSOM 

Chapter 16   Discussion: Biological Assumptions and Predictions
16.1     Local microcircuit for lateral interactions 

Chapter 17   Future Work: Computational Directions
17.1     High-level influence on illusory contour perception 
17.2     Example Topographica model 
17.3     Example Topographica screenshot 

Chapter A   LISSOM Simulation Specifications
A.1     Mapping between neural sheets in LISSOM