Exam 1 Flashcards
Key features of rate codes
*
spike rate encodes stimulus intensity
*
spike rate decreases with time (adaptation)
*
stimulus identityis determined by whichcell is firing (labelled line)
*
simple changes in response properties can enable coding of different stimulus features
*
selectivity arises from the biophysics of the sensory organ combined with intrinsic properties of the sensory neuron.
*
different neurons can represent the same stimulus differently.
Population coding: fundamental principles
*
combining activity across neurons expands coding ‘capacity’
*
in many (most!) cases, population coding is forced by biology
*
population activity vectors can reliably encode multi-parameter features
*.synchronous spiking may link assemblies of neurons together for binding features into objects
Challenge: Devise a neural circuit that is sensitive to first spikes or to a specific order of inputs.
Feed-forward excitation
Convergent FFE: Neurons respond selectively to certain features (not seen by any one of upstream/presynaptic neurons). Increases signal to noise ratio. Integration/summation for selectivity. Coincidence detection (“and” operator)
Divergent: allows signal to be processed by multiple downstream areas. (dimensionality expansion)
MAMMALIAN VISUAL SYSTEM
Feed-forward inhibition
Feed-back inhibition
both are used to regulate
the duration and magnitude of incoming
excitatory signals.
such as regulating
the gain and dynamic range of input
signals and facilitating synchronous or oscillatory
firing
FF_i
E-I balance
Fast Temporal integration window
Gain modulation (either divisive or subtractive depending on how inhibition is structured on the I side)
Almost every excitatory connection
in the visual pathway is accompanied
by feedforward inhibition, feedback
inhibition, or both.
FB_i :
Gain modualation on the O
Temporal window with longer time window due to the synaptic delay
Lateral Inhibition
It selects information to be
propagated to downstream circuits by amplifying
differences in activity between parallel
pathways.
-lateral inhibition is used to enhance selectivity and contrast. Example center-surround inhibition.
-winner-take-all
-attractor dynamics
ex. Odor coding / grid cells also responsible for center on, surround off visual receptive fields.
mutual inhibition
-oscillations/synchrony.
ex. locomotoin, Stomatogastric Nervous System of Lobsters
Marr Levels
Level 1: Computation. Why (problem)
Level 2: Algorithm. What (rules)
Level 3: Implementation. How (physical)
topographic mapping
Neighboring
input neurons connect to neighboring target
neurons through orderly axonal projections
ex. retinotopy. tonotopy. sensorimotor
Discrete parallel processing
allows
signals to be represented and processed in
parallel by discrete information channels. Eventually converge which increases SNR
ex. odor
Dimensionality Expansion
relatively small number of inputs) synapse
onto a much larger number of output neurons.
allowing output
neurons to represent distinct combinations
of inputs.
rheobase
: the minimal electric current required to excite a tissue (as nerve or muscle) given an indefinitely long time during which the current is applied
Mechanisms of Gain Modulation
variations of statistics of synaptic inputs
Synaptic input regulation:
- fluctuations in membrane potential
- changes in conductance state of cell/depolarization
Main takeaway from Otchy et al.
the manipulations affected independent functions of downstream circuits!
Main takeaway from Krakauer
Neurons observed during behavior don’t necessarily produce behavior/may not be the mechanism
Lien Paper
Q_thal untuned to orientation/direction, but F1 was very tuned. Q_thal not being orientation selective indicates that orientation selectivity doesn’t arise from convergence of thalamic neurons whose firing rates are orientation selective.
Q_sub tuned to orientation/direction, therefore Q_total is as well. F1_sub even more tuned to orientation/direction. F1_sub and Q_sub were tuned to same orientation and similar directions.
F1_thal and F1_sub have similar tuning and are nearly in-phase
orientation selectivity of thalamic neurons is a result of the centers of their respective on/off RFs. the angle between them allows them to be tuned to orientation.
