Note: Grayscal threshold in the simple method is 155 as 0 represents white and 255 represents black in this page.
Grayscale threshold would be (255-155)= 100 if "grayscale" representation is reversed.

Table 0. Comparison of recognition methods Image Ctrl Case: 458 AD Case: 311 Simple method(threshod:155, size cutoff:10) 73.7% ( 9.4% ) 74.7% ( 7.4% ) Potts Model with 1 condition 85.9% ( 5.2% ) 77.4% ( 4.3% ) Potts Model with 4x4 conditions with Kappa=10 97.8% ( 3.0% ) 93.4% ( 2.1% ) Potts Model with 4x4 conditions with kappa=0 95.0% ( 4.9% ) 93.1% ( 2.1% )

Table 1. Neuron recognition percentage(4x4 sets of conditions) Image C458-a2 C458-c2 C458-c4 C458-a7 C458-d2 Over-all for 5 Number of neurons(n1) 541 401 588 135 479 2144 Number of clusters(n2) 539 401 592 134 495 2161 Correctly recognized number(n3) 529 392 576 130 470 2097 Recognization Percentage (n3/n1) 97.8% 97.8% 98.0% 96.3% 98.1% 97.8% Percentage of extra clusters(1-n3/n2) 1.9% 2.2% 2.7% 3.0% 5.0% 3.0%

Table 2. Neuron recognition percentage(One set of condition: T=0.4,D=2.5,K=10) Image C458-a2 C458-c2 C458-c4 C458-a7 C458-d2 Over-all for 5 Number of neurons(n1) 497 369 531 107 398 1902 Number of clusters(n2) 458 332 477 73 384 1724 Correctly recognized number(n3) 434 313 459 69 359 1634 Recognization Percentage (n3/n1) 87.3% 84.8% 86.4% 64.5% 90.2% 85.9% Percentage of extra clusters(1-n3/n2) 5.2% 2.7% 3.8% 5.5% 6.5% 5.2%

Table 3. Neuron recognition percentage (Grayscale threshold is 155 and size cutoff is 10) Image C458-a2 C458-c2 C458-c4 C458-a7 C458-d2 Over-all for 5 Number of neurons(n1) 473 361 508 117 342 1801 Number of clusters(n2) 378 312 392 105 278 1465 Correctly recognized number(n3) 334 289 359 101 245 1328 Recognization Percentage (n3/n1) 70.6% 80.1% 70.7% 86.3% 71.6% 73.7 % Percentage of extra clusters(1-n3/n2) 13.2% 7.4% 8.4% 3.8% 11.9% 9.4%

Table 1. Neuron recognition percentage(4x4 sets of conditions) Image AD311-b3 AD311-b4 AD311-c3 AD311-c4 AD311-d4 Over-all for 5 Number of neurons(n1) 575 281 341 552 303 2052 Number of clusters(n2) 554 266 331 524 284 1959 Correctly recognized number(n3) 548 262 317 516 274 1917 Recognization Percentage (n3/n1) 95.3% 93.2% 93.0% 93.5% 90.0% 93.4% Percentage of extra clusters(1-n3/n2) 1.1% 1.5% 4.2% 1.5% 3.5% 2.1%

Table 2. Neuron recognition percentage(One set of condition: T=0.4,D=2.5,K=10) Image AD311-b3 AD311-b4 AD311-c3 AD311-c4 AD311-d4 Over-all for 5 Number of neurons(n1) 550 238 286 504 240 1818 Number of clusters(n2) 468 165 241 431 167 1472 Correctly recognized number(n3) 443 159 228 419 159 1408 Recognization Percentage (n3/n1) 80.5% 66.8% 79.7% 83.1% 66.3% 77.4% Percentage of extra clusters(1-n3/n2) 5.3% 3.6% 5.4% 2.8% 4.8% 4.3%

Table 3. Neuron recognition percentage (Grayscale threshold is 155 and size cutoff is 10) Image AD311-b3 AD311-b4 AD311-c3 AD311-c4 AD311-d4 Over-all for 5 Number of neurons(n1) 498 229 253 473 234 1687 Number of clusters(n2) 411 182 206 387 174 1360 Correctly recognized number(n3) 377 171 195 349 168 1260 Recognization Percentage (n3/n1) 75.7% 74.7% 77.1% 73.8% 71.8% 74.7% Percentage of extra clusters(1-n3/n2) 8.3% 6.0% 5.3% 9.8% 3.4% 7.4%