HOME | Adjacency vs. Wirelength | Macaque | Cat | Metamodules | Microchip | Macroeconomy | Conclusion | References

Table S6. Component placement optimization: results summary.  "Estimated Rank" designates proportion of all possible alternative layouts that are of lower connection-cost than the actual layout.  Each layout is scored for violations of the Adjacency Rule. Size Law goodness of fit r2 is for a model of the form of an inverse exponential function y = ke-mx.
       

Number of   
Components   

 Estimated      
Rank                 		 Size Law
r2

Neural Networks

Caenorhabditis elegans

11

 4.0 10-7 0.99
Macaque Visual Cortex

17

 1.2 10-7 0.91
Cat Visual Cortex

15

 7.2 10-6 0.94
Cat Cortex Metamodules

14

 1.1 10-7 0.97
Cat Cortex:
    Vis, Aud, Somato

39

 [(3x) 1011]* ---


Non-Neural Networks

AMI49 Microchip
    Esbenson & Kuh Layout

15

 7.0 10-2 0.77
    Hong et al Layout

15

 3.3 10-1 0.69
    Lin & Chang Layout

15

 1.5 10-3 0.78
Macroeconomic Networks
    USA Commodity-Flow

15

 --- ---
    Europe Ex/Im

8

 4.0 10-4 ---


*In each of 3 separate replications, 100 billion randomly sampled layouts were tested without finding a better layout than the actual one.


Table S7.  Connections vs adjacencies among network components: 2 x 2 contingency tables.  For each neural system, the relationship is highly significant: p < 0.0001; r > 0.30.  Connections and adjacencies to immediately contiguous edge-components are included.  (Number of core components is given in parentheses.)

Cortex Networks

Macaque Visual Cortex (17)
                Adj     NotAdj    Total    X2                r
Con 38 82 120 30.6433 0.30801
NotCon 16 187 203
Total 54 269 323

Cat Visual Cortex (15)
Adj NotAdj Total X2 r
Con 40 86 126 40.0761 0.365496
NotCon 8 166 174
Total 48 252 300

Cat Metamodules (14)
Adj NotAdj Total X2 r
Con 44 90 134 37.09234 0.368605
NotCon 6 133 139
Total 50 223 273

Cat Vis/Aud/Som Cortex (39)
Adj NotAdj Total X2 r
Con 101 350 451 143.0408 0.333381
NotCon 18 818 836
Total 119 1168 1287

Non-Neural Networks

AMI49 Microchip

Esbensen & Kuh Layout (15)
Adj NotAdj Total X2 r
Con 17 95 112 0.62703 0.0409
NotCon 32 231 263
Total 49 326 375

Hong et al Layout (15)
Adj NotAdj Total X2 r
Con 14 86 100 0.00072 0.0016
NoCon 24 146 170
Total 38 232 270

Lin & Chang Layout (15)
Adj NotAdj Total X2 r
Con 19 84 103 6.62842 0.1451
NotCon 18 194 212
Total 37 278 315

Macroeconomic Networks

USA Commodity-Flow (15) (@1500)
Adj NotAdj Total X2 r
Con 47 137 184 29.4310924 0.274708
NotCon 12 194 206
Total 59 331 390

Europe Ex/Im (8) (@1250)
Adj NotAdj Total X2 r
Con 23 36 59 8.06358 0.3098
NotCon 2 23 25
Total 25 59 84


Previous | Next



Copyright Christopher Cherniak (2002)
Last Modified 11/19/2002