|Author||Randall C. O'Reilly|
|First Published||Jul 26 2016|
|Tags||Inhibition, Feedforward, Feedback, GABA, Activity Regulation, Competition, kWTA|
|Description||Inhibitory interneurons control the activation of excitatory pyramidal neurons, through both feedforward and feedback projections. The result is a k-winners-take-all distributed pattern of activity.|
|Updated||28 July 2016, 6 September 2016, 13 January 2017, 11 January 2018, 6 February 2018|
|Versions||8.0.0, 8.0.2, 8.0.3, 8.0.4, 8.0.5|
|Emergent Versions||8.0.0, 8.0.4, 8.5.0, 8.5.1|
This simulation explores how inhibitory interneurons can dynamically control overall activity levels within the network, by providing both feedforward and feedback inhibition to excitatory pyramidal neurons. This inhibition is critical when neurons have bidirectional excitatory connections, as otherwise the positive feedback loops will result in the equivalent of epileptic seizures -- runaway excitatory activity.
The network in the right view panel contains a 10x10 unit input layer, which projects to both the 10x10 hidden layer of excitatory units, and a layer of 20 inhibitory neurons. These inhibitory neurons will regulate the activation level of the hidden layer units, and should be thought of as the inhibitory units for the hidden layer (even though they are in their own layer for the purposes of this simulation). The ratio of 20 inhibitory units to 120 total hidden units (17 percent) is like that found in the cortex, which is commonly cited as roughly 15 percent (White, 1989a; Zilles, 1990). The inhibitory neurons are just like the excitatory neurons, except that their outputs contribute to the inhibitory conductance of a neuron instead of its excitatory conductance. We have also set one of the activation parameters to be different for these inhibitory neurons, as discussed below.
Let's begin as usual by viewing the weights of the network.
| ⇒ Select r.wt in the netview control panel and then click on some of the hidden layer and inhib layer units with the red arrow. |
Most of the weights are random, except for those from the inhibitory units, which are fixed at a constant value of .5. Notice also that the hidden layer excitatory units receive from the input and inhibitory units, while the inhibitory units receive feedforward connections from the input layer, and feedback connections from the excitatory hidden units, as well as inhibitory connections from themselves.
Now, we will run the network. Note the graph view above the network, which will record the overall levels of activation (average activation) in the hidden and inhibitory units.
| ⇒ Select act to view activations in the network window, and press and in the . |
You will see the input units activated by a random activity pattern, and after several cycles of activation updating, the hidden and inhibitory units will become active. The activation appears quite controlled, as the inhibition counterbalances the excitation from the input layer. From the average activity plotted in the graph above the network, you should see that the hidden layer (black line) has around 20 percent activation.
In the next sections, we manipulate some of the parameters in the control panel to get a better sense of the principles underlying the inhibitory dynamics in the network. Because we will be running the network many times, you may want to toggle the network display off to speed up the settling process if your computer is running too slowly (the graph log contains the relevant information anyway).
| ⇒ To toggle the display off, click the Display button in the upper left-hand corner of the network view panel window. |
Strength of Inhibitory Conductances
Let's start by manipulating the maximal conductance for the inhibitory current into the excitatory units, ff_hidden_g_bar.i, which multiplies the level of inhibition coming into the hidden layer (excitatory) neurons. Clearly, one would predict that this plays an important role.
| ⇒ Decrease ff_hidden_g_bar.i from .4 to .3 and press . Then increase it to .5 and press . |
|Question 3.6: What effects does decreasing and increasing ff_hidden_g_bar_i have on the average level of excitation of the hidden units and of the inhibitory units, and why does it have these effects?|
| ⇒ Set ff_hidden_g_bar.i back to .4 (or just hit the button. |
Now, let's see what happens when we manipulate the corresponding parameter for the inhibition coming into the inhibitory neurons, ff_inhib_g_bar.i. You might expect to get results similar to those just obtained for ff_hidden_g_bar.i, but be careful -- inhibition upon inhibitory neurons could have interesting consequences.
| ⇒ First ff_inhib_g_bar.i of .75 for comparison. Then decrease ff_inhib_g_bar.i to .6 and , and next increase ff_inhib_g_bar.i to 1.0 and . and with a |
With a ff_inhib_g_bar.i of .6, you should see that the excitatory activation drops, but the inhibitory level stays roughly the same! With a value of 1.0, the excitatory activation level increases, but the inhibition again remains the same. This is a difficult phenomenon to understand, but the following provide a few ways of thinking about what is going on.
First, it seems straightforward that reducing the amount of inhibition on the inhibitory neurons should result in more activation of the inhibitory neurons. If you just look at the very first blip of activity for the inhibitory neurons, this is true (as is the converse that increasing the inhibition results in lower activation). However, once the feedback inhibition starts to kick in as the hidden units become active, the inhibitory activity returns to the same level for all runs. This makes sense if the greater activation of the inhibitory units for the ff_inhib_g_bar.i = .6 case then inhibits the hidden units more (which it does, causing them to have lower activation), which then would result in less activation of the inhibitory units coming from the feedback from the hidden units. This reduced activation of the inhibitory neurons cancels out the increased activation from the lower ff_inhib_g_bar.i value, resulting in the same inhibitory activation level. The mystery is why the hidden units remain at their lower activation levels once the inhibition goes back to its original activation level.
One way we can explain this is by noting that this is a dynamic system, not a static balance of excitation and inhibition. Every time the excitatory hidden units start to get a little bit more active, they in turn activate the inhibitory units more easily (because they are less apt to inhibit themselves), which in turn provides just enough extra inhibition to offset the advance of the hidden units. This battle is effectively played out at the level of the derivatives (changes) in activations in the two pools of units, not their absolute levels, which would explain why we cannot really see much evidence of it by looking at only these absolute levels.
A more intuitive (but somewhat inaccurate in the details) way of understanding the effect of inhibition on inhibitory neurons is in terms of the location of the thermostat relative to the AC output vent -- if you place the thermostat very close to the AC vent (while you are sitting some constant distance away from the vent), you will be warmer than if the thermostat was far away from the AC output. Thus, how strongly the thermostat is driven by the AC output vent is analogous to the ff_inhib_g_bar.i parameter -- larger values of ff_inhib_g_bar.i are like having the thermostat closer to the vent, and will result in higher levels of activation (greater warmth) in the hidden layer, and the converse for smaller values.
| ⇒ Set ff_inhib_g_bar.i back to .75 before continuing (or hit ). |
Roles of Feedforward and Feedback Inhibition
Next we assess the importance and properties of the feedforward versus feedback inhibitory projections by manipulating their relative strengths. The control panel has two parameters that determine the relative contribution of the feedforward and feedback inhibitory pathways: ff_ff_inhib_wt_scale_rel applies to the feedforward weights from the input to the inhibitory units, and ff_fb_inhib_wt_scale_rel applies to the feedback weights from the hidden layer to the inhibitory units. These parameters (specifically the .rel components) uniformly scale the strengths of an entire projection of connections from one layer to another, and are the arbitrary wt_scale.rel (r_k) relative scaling parameters described in Net Input Detail from the Neuron Chapter.
| ⇒ Set ff_ff_inhib_wt_scale_rel to 0, effectively eliminating the feedforward excitatory inputs to the inhibitory neurons from the input layer (i.e., eliminating feedforward inhibition). |
|Question 3.7: How does eliminating feedforward inhibition affect the behavior of the excitatory and inhibitory average activity levels -- is there a clear qualitative difference in terms of when the two layers start to get active, and in their overall patterns of activity, compared to with the default parameters?|
| ⇒ Next, set ff_ff_inhib_wt_scale_rel back to 1 and set ff_fb_inhib_wt_scale_rel to 0 to turn off the feedback inhibition, and . |
Due to the relative renormalization property of these .rel parameters, you should see that the same overall level of inhibitory activity is achieved, but it now happens quickly in a feedforward way, which then clamps down on the excitatory units from the start -- they rise very slowly (and eventually do achieve the similar levels as before, but we don't run long enough to see that).
These exercises should help you to see that a combination of both feedforward and feedback inhibition works better than either alone, for clear principled reasons. Feedforward can anticipate incoming activity levels, but it requires a very precise balance that is both slow and brittle. Feedback inhibition can react automatically to different activity levels, and is thus more robust, but it is also purely reactive, and thus can be unstable and oscillatory unless coupled with feedforward inhibition.
Time Constants and Feedforward Anticipation
We just saw that feedforward inhibition is important for anticipating and offsetting the excitation coming from the inputs to the hidden layer. In addition to this feedforward inhibitory connectivity, the anticipatory effect depends on a difference between excitatory and inhibitory neurons in their rate of updating, which is controlled by the dt.net_tau parameters ff_hidden_dt_net_tau and ff_inhib_dt_net_tau in the control panel (see CCNBook/Neuron#Neural_Integration). As you can see, the excitatory neurons are updated at tau of 40 (slower), while the inhibitory are at 20 (faster) -- these numbers correspond roughly to how many cycles it takes for a substantial amount of change happen. The faster updating of the inhibitory neurons allows them to more quickly become activated by the feedforward input, and send anticipatory inhibition to the excitatory hidden units before they actually get activated.
| ⇒ To verify this, click on Defaults, set ff_inhib_dt_net_tau to 40 (instead of the 20 default), and then . |
You should see that the inhibition is no longer capable of anticipating the excitation as well, resulting in larger initial oscillations. Also, the faster inhibitory time constant enables inhibition to more rapidly adapt to changes in the overall excitation level. There is ample evidence that cortical inhibitory neurons respond faster to inputs than pyramidal neurons (e.g., Douglas & Martin, 1990).
One other important practical point about these update rate constants will prove to be an important advantage of the simplified inhibitory functions described in the next section. These rate constants must be set to be relatively slow to prevent oscillatory behavior.
| ⇒ To see this, press ff_inhib_dt_net_tau to 2.5, and ff_hidden_dt_net_tau to 5 and . , and then set |
These oscillations are largely prevented with finer time scale upgrading, because the excitatory neurons update their activity in smaller steps, to which the inhibitory neurons are better able to smoothly react.
Effects of Learning
One of the important things that inhibition must do is to compensate adequately for the changes in weight values that accompany learning. Typically, as units learn, they develop greater levels of variance in the amount of excitatory input received from the input patterns, with some patterns providing strong excitation to a given unit and others producing less. This is a natural result of the specialization of units for representing (detecting) some things and not others. We can test whether the current inhibitory mechanism adequately handles these changes by simulating the effects of learning, by giving units excitatory weight values with a higher level of variance.
| ⇒ First, press to return to the default parameters. this case to get a baseline for comparison. |
In this case, the network's weights are produced by generating random numbers with a mean of .25, and a uniform variance around that mean of .2.
| ⇒ Next, set the wt_type parameter in the control panel to TRAINED, and hit the button. |
The weights are then initialized with the same mean but a variance of .7 using Gaussian (normally) distributed values. This produces a much higher variance of excitatory net inputs for units in the hidden layer. There is also an increase in the total overall weight strength with the increase in variance because there is more room for larger weights above the .25 mean, but not much more below it.
| ⇒ Press to see what difference this makes for the overall excitatory level. |
You should observe a greater level of excitation using the trained weights compared to the initial untrained weights.
| ⇒ You can verify that the system can compensate for this change by increasing the ff_hidden_g_bar_i to .5. |
| ⇒ Before continuing, set the wt_type parameter in the control panel back to UNTRAINED, and hit the button. |
To make things simpler at the outset, we have so far been exploring a relatively easy case for inhibition where the network does not have bidirectional excitatory connectivity, which is where inhibition really becomes essential to prevent runaway positive feedback dynamics. Now, let's try running a network with two bidirectionally connected hidden layers.
| ⇒ First, select network in the control panel to BidirExcite instead of FFExcite. Then click on the tab in the right viewer panel to view this network. to get back the default parameters, do a for comparison, and then set |
In extending the network to the bidirectional case, we also have to extend our notions of what feedforward inhibition is. In general, the role of feedforward inhibition is to anticipate and counterbalance the level of excitatory input coming into a layer. Thus, in a network with bidirectional excitatory connectivity, the inhibitory neurons for a given layer also have to receive the top-down excitatory connections, which play the role of "feedforward" inhibition.
| ⇒ Verify that this network has both bidirectional excitatory connectivity and the "feedforward" inhibition coming back from the second hidden layer by examining the r.wt weights as usual. |
| ⇒ Now this network. |
The graph view shows the average activity for only the first hidden and inhibitory layers (as before). Note that the initial part up until the point where the second hidden layer begins to be active is the same as before, but as the second layer activates, it feeds back to the first layer inhibitory neurons, which become more active, as do the excitatory neurons. However, the overall activity level remains quite under control and not substantially different than before. Thus, the inhibition is able to keep the positive feedback dynamics fully in check.
Next, we will see that inhibition is differentially important for bidirectionally connected networks.
| ⇒ Set the bidir_hidden_g_bar.i parameter (note the bidir -- in the 2nd group of parameters) to .35, and . |
This reduces the amount of inhibition on the excitatory neurons. Note that this has a relatively small impact on the initial, feedforward portion of the activity curve, but when the second hidden layer becomes active, the network becomes catastrophically over activated -- an epileptic fit!
| ⇒ Set the bidir_hidden_g_bar.i parameter back to .4. |
Exploration of FFFB Inhibition
You should run this section after having read the CCNBook/Networks#FFFB_Inhibition_Function section of the main text.
| ⇒ Reset the parameters to their default values using the network to BidirExcite, and to get the initial state of the network. button, set the |
This should reproduce the standard activation graph for the case with actual inhibitory neurons.
| ⇒ Now, set the bidir_lay_inhib.on ON to use the FFFB function described in the main text. Also set the bdir_hidden_g_bar.i and bdir_inhib_g_bar.i parameters to 1.0 (otherwise the computed inhibition will be inaccurate), and the rate constant parameters to their defaults for normal (non unit inhib) operation, which is bdir_hidden_dt_net_tau and bdir_inhib_dt_net_tau both to 1.4. Finally, you need to turn off the inhibitory projections (when present, these will contribute in addition to whatever is computed by FFFB) -- set bdir_inhib_wt_scale.abs to 0 (this sets the absolute scaling factor to 0, effectively nullifying these connections). Hit Apply. |
| ⇒ Press . |
The activations should be right around the 10-15% activity level. How does this change with trained weights as compared to the default untrained weights?
| ⇒ Set wt_value to TRAINED and do , and then again. |
You should see the hidden activities approach the 20% level now -- this shows that FFFB inhibition is relatively flexible and overall activity levels are sensitive to overall input strength. You should also notice that FFFB dynamics allow the network to settle relatively quickly -- this is due to using direct and accurate statistics for the incoming netinput and average activations, as compared to the more approximate sampling available to interneurons. Thus, FFFB is probably still more powerful and effective than the real biological system, but this does allow us to run our models very efficiently -- for a small number of cycles per input.
| ⇒ To test the set point behavior of the FFFB functions, we can vary the amount of excitatory input by changing the input_pct levels, to 10 and 30 instead of the 20% default. After you change input_pct, you need to press the button (this makes a new input pattern with this percent of neurons active), and then . |
|Question 3.8: How much does the hidden average activity level vary as a function of the different input_pct levels (10, 20, 30). What does this reveal about the set point nature of the FFFB inhibition mechanism?|
Finally, you can explore the effects of changing the gi, ff, and fb parameters in the bidir_lay_inhib field, which change the overall amount of inhibition, and amounts of feedforward and feedback inhibition, respectively.
| ⇒ You may now close the project (use the window manager close button on the project window or menu item) and then open a new one, or just quit emergent entirely by doing menu option or clicking the close button on the root window. |