The plasticity direction differs between Up/Downbound micromodules.
<p><b>(A)</b> BCM mechanism: the x-axis represents the activity of the PC, and the y-axis represents the predicted plasticity change at a PF-PC synapse before scaling by PF activity. Positive values indicate potentiation; negative values indicate depression. Each gray curve shows t...
-д хадгалсан:
| Үндсэн зохиолч: | |
|---|---|
| Бусад зохиолчид: | , , , , , |
| Хэвлэсэн: |
2025
|
| Нөхцлүүд: | |
| Шошгууд: |
Шошго нэмэх
Шошго байхгүй, Энэхүү баримтыг шошголох эхний хүн болох!
|
| Тойм: | <p><b>(A)</b> BCM mechanism: the x-axis represents the activity of the PC, and the y-axis represents the predicted plasticity change at a PF-PC synapse before scaling by PF activity. Positive values indicate potentiation; negative values indicate depression. Each gray curve shows the BCM function for a different value of the plasticity threshold <i>. The colormap encodes the corresponding</i> <i>values (in Hz), with lighter gray indicating lower thresholds. As PC activity increases, the sliding threshold</i> <i>shifts rightward, reducing the potentiation range and increasing the chances of depression (a larger part of the curve is negative). The same applies for a lower activity in PC, leading to higher chance of potentiation.</i> <b>(B)</b> CSpk-triggered change in BCM. When there is a CSpk, the activity of the PC (blue) decreases. The sliding threshold <i>(black) follows and goes lower than the PC activity, leading to potentiation (green). When</i> <i>is higher than the activity there is depression (red). This is shown for both Up/Downbound zones.</i> <b>(C)</b> <i>Trace of the IO membrane potential for Downbound coupled scenario. This trace was selected to illustrate the short-timescale effects of different IO spike intervals on plasticity mechanisms (BCM and CSpk-triggered LTD). While this particular trace appears bursty, it is not representative of the full IO population. IO neurons in the model exhibit a range of firing modes including tonic firing and quiescence. The mean response profile shows a ~ 1 Hz firing rate (</i><b><a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1013609#pcbi.1013609.g001" target="_blank">Fig 1B</a></b><i>) and the IO has a large distribution of firing rates across the population (</i><b><a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1013609#pcbi.1013609.g003" target="_blank">Fig 3B</a></b><i>). IO burstiness was not explicitly tuned and remained within biologically observed ranges (<6Hz), consistent with experimental findings [</i><a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1013609#pcbi.1013609.ref052" target="_blank">52</a>,<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1013609#pcbi.1013609.ref077" target="_blank">77</a>–<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1013609#pcbi.1013609.ref080" target="_blank">80</a><i>].</i> <b>(D)</b> CSpk-triggered LTD component of synaptic plasticity. This shows the change in synaptic weight due to a IO spike (CSpk) event, proportional to the instantaneous activity of the PF. In the absence of a CSpk, this component is zero; following a CSpk, it produces a transient depression that decays back to zero. <b>(E)</b> BCM component of synaptic plasticity. This reflects the change in synaptic weight based on the recent activity of the PC and the BCM threshold <i>. It evolves continuously, independently of CF input.</i> <b>(F)</b> <i>Total synaptic weight update, computed as the sum of the BCM component and the CSpk-triggered LTD component. All values are unitless and represent normalized synaptic efficacy changes per time step in the model.</i> <b>(G)</b> <i>Relative difference of synaptic weights of all PCs connecting to PF 1 before and after each plasticity (AP 1, 2, 3 and 4) between the epochs (current weight divided by the one of the previous event) for Upbound zones.</i> <b>(H)</b> <i>Same as I but for the Downbound zones.</i> <b>(I)</b> <i>Boxplots depict absolute differences between the epochs for the whole population of weights (current weight minus previous weight) for Upbound zones.</i> <b>(J)</b> <i>Same as K but for the Downbound zones.</i> <b>(K)</b> <i>Median synaptic weight changes across four plasticity epochs for 10 independent Upbound simulations with randomized network and OU input per run.</i> <b>(L)</b> <i>Same as M, but for Downbound simulations. Synaptic weights are unitless and represent normalized efficacy values in the model.</i></p> |
|---|