Announcement

Collapse
No announcement yet.

NVIDIA Announces "Xavier" AI Supercomputer SoC With Custom ARM64 + Volta GPU

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #21
    Originally posted by starshipeleven View Post
    Bullshit, all neurons by definition do that.
    No, many neurons, especially in the brain, do not fire action potentials. As you seem to like wikipedia: https://en.wikipedia.org/wiki/Non-spiking_neuron.

    Originally posted by starshipeleven View Post
    I suggest you read up about the Action Potential. https://en.wikipedia.org/wiki/Action_potential
    I suggest you read it. Nowhere does it say action potentials alone are based on a sum, not to mention the entire neuron. On the contrary, it points out that action potentials are dependent on the interaction of multiple ions flowing in different directions (which mathematically is a system of partial differential equations that varies from neuron to neuron), and is highly dependent on the history of the neuron. No system designed to simulate real neurons uses a sum, the simplest model is an integrate and fire model, which reduces this system of differential equations to a single differential equation and a tacked-on kludge for the actual spikes.

    But that article only provides an extremely simplified look at just one factor. There is a lot more going on besides that. Once you finish actually reading the article you linked to, I suggest you read up on low-threshold-activated potassium channels, backwards-propagating action potentials, volume transmission, bursting behavior, onset neurons, and dendritic spines. And that is just the tip of the iceberg, that is a lot more going on besides that.

    Overall, it might be better not to presume to lecture a neuroscientist about the most basic aspects of neuroscience.

    Comment


    • #22
      Originally posted by TheBlackCat View Post
      No, many neurons, especially in the brain, do not fire action potentials. As you seem to like wikipedia: https://en.wikipedia.org/wiki/Non-spiking_neuron.
      Well, that's still using membrane potential differences to decide when fire, just not big spikes of potential like spiking neurons. The general mechanism is the same, only the triggers are different.

      Nowhere does it say action potentials alone are based on a sum, not to mention the entire neuron.
      In physiology, an action potential is a short-lasting event in which the electrical membrane potential of a cell rapidly rises and falls, following a consistent trajectory.

      If you open membrane potential https://en.wikipedia.org/wiki/Membrane_potential
      and read first sentence:
      Membrane potential (also transmembrane potential or membrane voltage) is the difference in electric potential between the interior and the exterior of a biological cell.

      But also if you read a bit more it's also stated below in the same page

      All cells in animal body tissues are electrically polarized – in other words, they maintain a voltage difference across the cell's plasma membrane, known as the membrane potential.

      On the contrary, it points out that action potentials are dependent on the interaction of multiple ions flowing in different directions (which mathematically is a system of partial differential equations that varies from neuron to neuron), and is highly dependent on the history of the neuron.
      I'm trying hard to see that you are saying "well, it's more complex than that" (because yes it is, It's not my field but I know enough), instead of random technobabble to convince fools that you are right.

      No system designed to simulate real neurons uses a sum
      I'm talking of "sum of stimuli" goddamnit, I'm not talking of voltages or capacitance in a 100% accurate description of how stuff works at low level. I'm talking high school biology at most, not university physics. Don't get me wrong.

      The synapses from other neurons open ion channels or activate ion pumps (i.e. depolarization, or iperpolarization), and there are various other chemical signals that also affect the speed at which ion channels open or at which ion pumps operate.

      The end result is that when the membrane potential drops below X, the cell fires. (same for the non-spiking neurons, just a smaller difference)

      Sure there is a circus of other funky stuff modulating and doing whatever but the core operation is what I said. I have no intention (or ability) to make a 300-page manual about neuroscience to throw after people in a forum.

      Overall, it might be better not to presume to lecture a neuroscientist about the most basic aspects of neuroscience.
      Overall, it might be better if neuroscientists were able to tell dumbed down information for mass-consumption from actual scientific claims, it might also be better if they didn't show lack of understanding of high-school biology (that thing about membrane potential above), and last but not least, it might be better if neuroscientists were able to provide something more neuroscience-y than wikipedia to back their claims.

      Comment


      • #23
        damnit, unapproved post for TheBlackCat above

        Comment

        Working...
        X