THINKING BACTERIAL OR NEURONAL STYLE: BEN JACOB?

°
DID YOU SEE THE GOOGLE ADSENSE ADS ON THE RIGHT?

1) A REMINDER ON BACTERIA, NEURONS AND THINKING
----------------------------------------------------------------------

As you know from my first post, my novel entitled"WE SHR", a mispelling for "WE SHARE", presents the adventures of what I call Bactorgs, bacterial colonies having social and intelligent behaviors (e.g. Weiss's work). It also present Andrones, cultured neural networks showing also social and intelligent behaviors (e.g. Potter's work). Finally, and "audaciously", the novel will discuss scientific theories about and computational models of these systems.

Why "audaciously": because there is a rule in novel writing; every time you insert an equation or a hard scientific reasoning, your audience decreases by 50%¨. By that account, the "WE SHARE" audience should be at the 0.00000001 level or perhaps worse. My goal is to make science pleasant. We'll see if I succeed.

So, we need to discuss thinking in bacteria, thinking in cultured neuronal networks and models of these systems. Unexpected as it once seemed to me, there is a researcher working, at a very high level, on these three subjects. Meet Eshel Ben Jacob. We will first discuss its work on bacteria. Then we'll see what he did on neurons and, later on in another post, his modeling work.





2) BEN JACOB'S WORK ON BACTERIAL COLONIES (MY BACTORGS)
--------------------------------------------------------------------------------

Seen from my point of view, Eshel is a first class "Jack of all trades". He is at Tel Aviv University where he does world class work in Physics and Biophysics. He is a deep thinker and works on many subjects. As you may guess, the ones which interest us at the moment are bacterial cultures.

• Eshel demonstrated that , when challenged to grow in difficult conditions, bacterial colonies respond by adopting very sophisticated spatial growing patterns optimizing their food intake. He challenges them to grow in conditions of limited nutrients or against antibiotics or with spatial obstacles and they respond by devising clever growing spatial patterns (mostly spatially organized fractals). Probably, bacterial colonies need to do some colony-level computation to do that.
  
• Here is the title of a paper, Eshel published in the journal "Trends in microbiology" . This paper was the most downloaded paper from that journal in 2004. You'll find it on Eshel's site (see later). Read it and think twice: It's not Sci Fi but hard, respectable and respected science

Can you believe this title? Eshel argues that his experiments and those of other show that bacteria have developed sophisticated communication capabilities (quorum sensing, chemicals and plasmid exchanges) which they use to reach a high level linguistic communication ability between themselves (shared interpretation of messages, dialogs and meaning-based communication allowing intentional (??) behavior, decision making, recognition and identification of other colonies, ... all Eshel's own words).

This may be called "bacterial social intelligence..." what else?

How can I extrapolate for "WE SHARE"?

It seems to me that the development of Eshel's idea requires going beyond communication to move into the realm of inheritable colonial memory and commonly shared genomic context. What I need is a computer or a Turing machine based on a bacterial colony (see later my discussion of issues like collective brains, fluid neural networks and amorphous computing).

Eshel Goes further. I do not understand him yet but he seems to imply that the limitations of Turing machines may be overcome by bacterial colonies. I do not claim to understand this or even to admit that they are limitations to Turing machines?

Here is a picture of two of Eshel's colonies. He argues that such fractal shapes cannot be generated by chance. He also suggests that they express some computational parallel algorithm at work. I agree, clearly, Mandelbrot is at work here. (With permission from Eshel Ben Jacob)





Remark however, that generating such shapes might, computationally speaking, be easier than expected and may need only much simpler algorithms than what you would think at first sight. For instance, I have myself shown that a drop of glue squeezed between two plates of glass which are then separated slowly, form on each plate a fractal tree which is determined not by intricate algorithms but simply by the speed of separation of the plates and the surface tension, adherence and flow properties of the glue. No complex algorithm is involved to generate this kind of complex fractal patterns but only a few parameters and some rules and physical laws.

Issues like Gellman's complexity, crypticity and depth are playing a role here (estimating the complexity of the algorithms needed to coordinate such a growth). That's food enough for another post.

I will soon re-read Eshel's papers. This is not always easy going and I'll write some more posts about it later. Anyway, here is a cornerstone of WE SHARE: a respectable scientific paper discussing bacterial intelligence, social life and colonial memory. Those are the three pillars on which, in my novel, I build my underground bacterial colonies, the Bactorgs. A bit of extrapolation and you have great perspectives. Actually, you do not have to really extrapolate. Eshel's paper is better than Sci Fi.

Here is a link to this paper on Eshel's site: Ben jacob's Trends in microbiology 2004 paper.
And here is a link To Eshel's site itself: Eshel's site.

3) BEN JACOB'S WORK ON CULTURED NEURONS NETWORKS (MY ANDRONES):
-----------------------------------------------------------------------------------------------

Independently of his work on bacteria, Eshel works also on multi-electrodes arrays and cultured networks (remember Potter's work discussed in my first post). So, he is a precursor, not only of bactorgs but also of Andrones.

His work is complementary to Potter's work. Potter has extremely sophisticated, closed loop, two way electrical communication and learning in a long term neuronal culture but he has no way to communicate chemically with his cultures.

That might be a problem since a lot of reward and penalty in the brain is probably done chemically (neuro-transmitters, endorphins, cannabinoids ...). Eshel developed a way to have his cultured network on the electrode array stimulated at selected points by chosen chemicals. Every year people discover new chemical neurotransmitters and find new roles for them, see for instance the recent discoveries on NO and now the new hypothesis on the role of chimiokines.

There is clearly now a need to combine Potter's and Ben Jacob's work.

Here is a picture which tells you all you need to know for now about Eshel's work.


In the upper panel, you see an MEA and its cultured network (1). Then a micropipet (2) used for local chemical stimulation at selected locations (probably mounted on some kind of stereotactic equipment. Potentials are recorded and sent to a computer (3). The volume of the stimulating droplets are controlled by a second micrometer and a syringe.

In the lower panel, we show the MEA (dots and black lines), its cultured network (cells on the blue plate) and on top, the translucent micropipet placed above one of the recording electrodes. For stimulation, Ben Jacob currently uses picrotoxin (an inhibitory antagonist, i.e. a substance relieving the inhibition of excitation felt by some neurons) and show that this stimulation makes it possible to imprint multiple memories in the form of collective modes of neuronal firing and that these memories persist for days. Memories.... an important element of what Andrones need! (With permission from Eshel Ben JAcob)

4) EXTRAPOLATING POTTER'S AND BEN JACOB'S WORK FOR WE SHR
------------------------------------------------------------------------------------

Let us consider the implications of coupling Potter's and Ben Jacob's works: chemical and electrical stimulation, long term culture, two way electrical communication for adaptive embodied behavior. Let's extrapolate a little bit.

Why not add some neurogenesis ability to cultured neuronal cultures (controlling stem cells is another hobby of Potter)? Why not also control and direct axonal growth?

Why not have multi-culture networks on the same or on different interconnected plates. A bit of cortical cells, a bit of hippothalamus, a bit of hippocampal and so on, all and each growing their own kind of connections and interconnecting between themselves?

(Note; there is a paper which I am, right now, unable to quote exactly which describes the guidance of axons by external stimulators along long distance (several millimeters ) lines. We could use such systems to interconnect separate subcultures together i.e. to build andrones as multi-cultures having specific global connection patterns.

What you get is then close to what I envision for what I call an "Androne", the neural culture hero in my novel. It lives on a set of MEA and Petri dishes and has all the features I've just listed + a few others.

NOTE: Just to tell you that, in the novel, the researcher developing Andrones has discovered a library of electrical signals and chemicals to communicate with them, make them learn, store and retrieve memories and so on... My Andrones will perhaps feel elementary emotions and perhaps some form of elementary consciousness (after all, it is probably a graded property). Is it too far fetched? What do you think? Compare with current models of pre-consciousness like Edelman's robots?

5) CONCLUSION
--------------------

All that to tell you that you have just met Eshel, one of my heroes. As Einstein said, you define a scientist by what he does. You have to admit Eshel does a lot and every bit of his work is minted in originality . Here we have just scratched the surface. Go diving deep in his site (see before)!

Again time to sleep, My cat Touti is meowing and that's the signal. In the next post, you'll meet some other people:

• Klaas Hellingwerf from Amsterdam who sees bacteria as proto-neurons and has organized the first European community sponsored workshop and research program on this theme.
• Bonnie Bassler who studies the languages bacteria use to speak between themselves ( in a given species and among species).

Together, they make an impressive case for the idea that bacteria have some elementary thinking ability (thinking being defined conservatively but operationally by what Mc Culloch and Pitts models can do..., a limited form of thinking admittedly but not a bad one as a starting point).

And as usual, a thought for the night:

To define art, don't ask what you can still add in a picture but what you can still erase...

Don't dream about Andrones and Bactorgs or, like me, just a little bit.

Jacques