Downloading SciLab

• SciLab is available from the web site http://www.scilab.org
• It will take around 500 MB on your computer. It also is available to use in the JMH 330 computer lab.

1. Open terminal
2. Type: cd Downloads/scilab/bin
3. Type: ./scilab

• Introductory SciLab commands

• Data for January 26 class class1.mat
• Data for homework 1 HW1data.mat
• S11_LOCbehav.mat — behavioral data for Subject S11
• fMRIdata1.mat — results of an MRI scan
• spikeExample.mat — sample spiking data
• hw3data.mat — rate and spiking data for HW 3
• S11_localizer.mat — neural localizer data
• convolving.mat — example patterns for convolution, to be discussed in class March 30

• forwardModelSim.sci — used to simulate the simplest motion feedback system, where the system begins at location 0 and moves to location targetLocation with variable delay in sensory feedback and with optional disturbance

  bodyPosition=forwardModelSim(targetLocation,speedMultiplier,sensoryDelay,simulationTime,disturbance);

  same input arguments as feedbackSim

• feedbackSim.sci — used to simulate the simplest motion feedback system, where the system begins at location 0 and moves to location targetLocation with variable delay in sensory feedback and with optional disturbance

  bodyPosition=feedbackSim(targetLocation,speedMultiplier,sensoryDelay,simulationTime,disturbance);

  speedMultiplier — at each time step, system speed is determined as speed = (desiredLocation-actualLocation)*speedMultiplier

  sensoryDelay — specified in positive integer time steps

  simulationTime — specifies the number of time steps over which to simulate the system

  disturbance — optional input; if provided, must be a vector with simulationTime number of entries specifying amount of disturbance at each time step

• pfcSim.sci — used to simulate neural dynamics, given an input value that changes over time, a memory value (1 for "remember" and 0 for "stop remembering") and a matrix of neuron-to-neuron weights

  neuralDynamics=pfcSim(feedforwardInput,memoryInput,weightMatrix);

• corr3250.sci — computes correlation between two vectors (will be discussed in class March 30)
• disp2d.sci — displays heat map of matrix data

  disp2d(matrix)       Display 2D matrix, with red being largest value and blue being smallest value in matrix

  disp2d(matrix,%F)   Display 2D matrix unscaled, with red being 1 and blue being 0

• disp2dB.sci — displays heat map of matrix data -- AXIS LABELS IN disp2dB ARE SOMEWHAT MORE INTUITIVE

  disp2dB(matrix)       Display 2D matrix, with red being largest value and blue being smallest value in matrix

  disp2dB(matrix,%F)   Display 2D matrix unscaled, with red being 1 and blue being 0

• spikeShow.sci — displays spikes for each neuron over time
• lifNeurons.sci — simulates "integrate and fire" neurons

  volts=lifNeurons(duration, numNeurons)	Returns voltage every 1ms over duration seconds for numNeurons simulated neurons given 100mV of input
  volts=lifNeurons(durations, numNeurons, opts)	Same as above, but opts variable alters simulation details

  opts.input	mV input to each neuron over time
  opts.wts	weight matrix among neurons (matrix of 0s by default))
  opts.Vreset	resting state voltage in mV (-70mv by default)
  opts.Vthresh	point (in mV) at which voltage is reset (10 mV by default)
  opts.tau	tau for neurons' voltage change (dv/dt)
  opts.tau2	tau for effects pre-synaptic spike on post-synaptic input voltage
  opts.dt	time step of simulation (.001 s/1 ms by default)