Stochastic resonance in a summing network with varied thresholds was investigated using GaAs-based etched nanowire field-effect transistors having different threshold voltages. The network's response adapted to input offset fluctuations in the range of the threshold voltage variation and the network could detect a weak signal without any adjustment of the input offset or the addition of high noise. The observed adaptability resulted from a widened dynamic range of the system due to signal decomposition and reconstruction by multiple thresholds together with the output summation process. © 2010 American Institute of Physics. ͓doi:10.1063/1.3428784͔The coexistence of noise and fluctuations created by utilizing stochastic resonance ͑SR͒ has recently been an interesting issue in electronics. 1-4 SR is a phenomenon in which the response of a system is enhanced by adding noise. 5,6 It is known to be a key mechanism for precise detection and transmission of weak signals comparable to the thermal energy in biological systems. 7 To date, many electronic systems that artificially cause SR have been investigated 8-12 but they have never been turned into practical applications excepting dithering. A major reason is that the SR at an optimal noise never improves the information content of the signal as that before adding noise. However it can improve the information transfer and the signal-to-noise ratio with noise, 13,14 and this feature is still useful for electronics. We have succeeded in causing SR in a semiconductor nanowire field-effect transistor ͑FET͒ and have demonstrated an enhancement of the response of an FET parallel summing network. 15 At this stage, the positive contribution of noise has been confirmed, but the influence of the physical fluctuation of network units on SR has not been clarified, although Collins et al. 16 suggested that the noise compensates for the variation in units leading to an optimally enhanced system response. In this Letter, we report on our investigation of the signal response of a summing network of GaAs-based nanowire FETs with threshold voltage variation. The threshold voltages of seven FETs were artificially modified by making devices of different sizes. A parallel summing network was formed using the different FETs and the system's SR response was characterized experimentally.An FET summing network with varied threshold voltages is schematically shown in Fig. 1͑a͒. The basic mechanism of SR is a noise-assisted state transition in a threshold system. 6,13 To cause SR, we utilized the nonlinearity of the FET gate threshold characteristic. 15 The input and output of this system were gate voltage V G and the sum of the sourcedrain current from each FET ͚I DS ͑=I out ͒, respectively. The FET was operated in the subthreshold region. The FETs were fabricated on nanowires formed by electron beam lithography and wet chemical etching of a conventional AlGaAs/ GaAs modulation-doped structure with a two-dimensional electron gas. A Schottky wrap gate ͑WPG͒ was formed on each nanowire. An exam...