"... In PAGE 4: ... In this way, with only 22 transistors for each cell, a su cient level of precision (8 bits of precision) was achieved. Table1 summarizes the main character- istics of the cortical module. 4 Conclusion We have demonstrated the analog VLSI feasibil- ity of complex resistive networks emulating cortical processing.... ..."

"... In PAGE 5: ... 2: Transitions between the two behavioral contexts Each behavioral context has its own set of mutual inhibition coefficients, which essentially specify the relative priorities of different behaviors. Table1 and Table 2 show these coefficients for the normal and emergency behavioral contexts respectively. The values of these coefficients are assigned based on the following heuristic rules: very strong inhibition (0.... In PAGE 5: ...4 0.4 x Casual walk Flee to exit Explore attractive point Maintain personal space Casual walk inhibiting inhibited Table1 and Table 2 show that the Flee to exit behavior has the highest relative priorities in both behavioral contexts, because it strongly inhibits other behaviors. As a result, this behavior plays a major role in the emergency context.... ..."

"... In PAGE 4: ...nput from the brainstem. The weights of the segmental connections are given in Table 1. The complete controller is formed by interconnecting 100 copies of this segmental network. An interconnection consists of extending the connection from one neuron to another in one segment to the corresponding neuron in neighbouring segments ( Table1 ). A neuron unit is modeled as a leaky integrator with a saturating transfer function.... In PAGE 4: ... The parameters of each type of neuron are given in Table 2. These parameters, as well as the connection weights of Table1 , have been de ned so that the simulation of the model ts physiological observations [Ekeberg 93]. When the segmental networks receive adequate excitation from the brainstem, they oscillate regularly with the left and right neurons out of phase.... In PAGE 20: ...0 [4, 1] - - 5.0 Table1 0: Evolved interconnections between biological segmental network (best of run2). The table gives the weights of the segmental network (identical to Ekeberg apos;s) and, between brackets, the extensions of the evolved interconnections in the rostral and caudal direction respectively.... In PAGE 20: ... These solutions have all similar intersegmental couplings and are signi cantly di er- ent from that of Ekeberg apos;s model in which only the CIN have asymmetric projections (favouring the caudal direction, cf. Table1 )(see Appendix F for a description of the evolved con gurations). Table 10 gives the interconnections of one of them.... In PAGE 20: ... The evolved controllers have interconnections which respect to some extent these observations. For the solution given in Table1 0, for instance, only the long rostral projections of the EIN to CIN, the CIN to LIN and CIN to MN strongly disagree with the biological observations. The others correspond to the observed biological interconnections if we hypothesize that the projections observed in the real lamprey from one type of neuron not necessarily tar- get all neuron types in other segments in an identical way.... In PAGE 30: ...7 - - 10.2 Table1 2: Evolved arti cial segmental oscillator, best of run1 0 500 1000 1500 2000 2500 3000 3500 4000 0 0.2 0.... In PAGE 31: ...3 11.3 Table1 3: Evolved arti cial segmental oscillator, best of run2 0 500 1000 1500 2000 2500 3000 3500 4000 0 0.2 0.... In PAGE 32: ...5 - - - - 7.6 Table1 4: Evolved arti cial segmental oscillator, best of run3 0 500 1000 1500 2000 2500 3000 3500 4000 0 0.2 0.... In PAGE 33: ...1 6.4 Table1 5: Evolved arti cial segmental oscillator, best of run4 0 500 1000 1500 2000 2500 3000 3500 4000 0 0.2 0.... In PAGE 34: ...9 - - 9.1 Table1 6: Evolved arti cial segmental oscillator, best of run5 0 500 1000 1500 2000 2500 3000 3500 4000 0 0.2 0.... In PAGE 35: ...0 - - 4.8 Table1 7: Evolved arti cial segmental oscillator, best of run6 0 500 1000 1500 2000 2500 3000 3500 4000 0 0.2 0.... In PAGE 36: ...2 - 3.8 Table1 8: Evolved arti cial segmental oscillator, best of run7 0 500 1000 1500 2000 2500 3000 3500 4000 0 0.2 0.... In PAGE 37: ...2 - - - - - 8.6 Table1 9: Evolved arti cial segmental oscillator, best of run8 0 500 1000 1500 2000 2500 3000 3500 4000 0 0.2 0.... ..."

"... In PAGE 5: ... be seen most clearly when comparing the architecture of these two subsystems. The strengths of the inhibitory connections are shown in Table3 . Refer to section 5.... ..."

"... In PAGE 2: ... Amantadine, an old antiviral compound, was also studied. Different terms have been used to express antiviral activi- ty, namely, EC50, 95% effective concentration (EC95), and 50% inhibitory concentration (IC50); Table1 illustrates the range of activity against selected viruses. Tenfold dilutions of the drug were tested to cover a broad range of concentrations above and below inhibitory dosages as reported by the manufacturer for other viral- host combinations.... ..."

"... In PAGE 3: ... vulgaris (Figure 2B, right). The inhibitory activi- ties of purified AtPGIP1 and AtPGIP2 were measured against several fungal PGs ( Table2 ). Both AtPGIPs exhibited com- parable inhibitory activity against PG of Botrytis but failed to inhibit PGs of A.... ..."

"... In PAGE 8: ... As a result, it will not wait any more and start to move forward. Table1 and Table 2 show the inhibitory coefficients of the team_forming context and the convoy context respectively. It can be seen that in the team_forming behavioral context, Search inhibits Wait with coefficient 0.... ..."