Biocomputing : the promise of biological computing

Advances in neu­ros­cience are paving the way for the construc­tion of bio­lo­gi­cal com­pu­ters using neu­rons and brain tis­sue. This inven­tion was both a break­through for com­pu­ting and a tool for fun­da­men­tal research and the stu­dy of human diseases.

They are some­times cal­led mini-brains. These cere­bral orga­noids are in fact 3‑dimensional cultures of cere­bral tis­sue. They could lead to the next revo­lu­tion in com­pu­ting, the birth of orga­noid intel­li­gence (OI).

As a remin­der, and contra­ry to what its name sug­gests, arti­fi­cial intel­li­gence (AI) does not mimic human intel­li­gence. In fact, it works in a com­ple­te­ly dif­ferent way. Just watch it play chess : it sacri­fices far more pieces than any human player¹. Ano­ther dif­fe­rence is its ener­gy consump­tion. In June 2022, the record capa­ci­ty of the Fron­tier super­com­pu­ter, hou­sed at Oak Ridge Natio­nal Labo­ra­to­ry in the Uni­ted States, was 1.1 exa­flops (i.e. it per­forms up to 1.1 1018 ope­ra­tions per second), a per­for­mance equi­va­lent to the human brain. Except that the human brain ope­rates on 20 Watts, whe­reas the Ame­ri­can super­com­pu­ter requires 10MW…

Orga­noid intel­li­gence (OI), on the other hand, pro­mises to bring the two sys­tems clo­ser toge­ther, by repro­du­cing the com­pu­ta­tio­nal per­for­mance of the best com­pu­ters and the ener­gy effi­cien­cy of a human brain. This tech­no­lo­gy is being deve­lo­ped at the cross­roads of three tech­no­lo­gi­cal break­throughs : elec­tro­phy­sio­lo­gy, arti­fi­cial intel­li­gence, and cere­bral organoids.

The first is essen­tial for com­mu­ni­ca­ting with cere­bral orga­noids. The chal­lenge ? Fin­ding a non-inva­sive sys­tem that takes into account the mul­tiple elec­tro­che­mi­cal signals that occur eve­ry second in the small mass of tis­sue in culture. Resear­chers at Ame­ri­can uni­ver­si­ties are pro­po­sing the use of cage-sha­ped elec­trodes², an ini­tial solu­tion for direct com­mu­ni­ca­tion with the orga­noid in culture.

The deve­lop­ment of the bio­lo­gi­cal com­pu­ter is still in its infancy.

The ques­tion of elec­trodes is also impor­tant for the deve­lop­ment of brain tis­sue. Without signals, cere­bral orga­noids can­not be construc­ted in three dimen­sions and remain inope­ra­tive. Nerve tis­sue has to work to func­tion and brain mecha­nisms require more than a simple elec­tro­che­mi­cal signal. Memo­ry involves the reor­ga­ni­sa­tion of neu­ron net­works and the inter­ven­tion of other brain cells, such as those that make up micro­glia, the immune sys­tem of the brain. To deve­lop, orga­noid intel­li­gence must inte­grate all these parameters.

The second tech­no­lo­gi­cal break­through that makes OI pos­sible is AI. It is essen­tial for explo­ring what these bio­lo­gi­cal sys­tems can do. Brain tis­sue cultures pro­duce very large volumes of data, both spa­tia­li­sed and struc­tu­red over time. Inter­pre­ting this data is a chal­lenge that recent advances in algo­rithms will be able to meet. 

The last tech­no­lo­gy on which the deve­lop­ment of OI is based has yet to prove itself, and that is the one that will enable brain orga­noids to change scale. At present, the lar­gest brain tis­sue cultures mea­sure just a few mil­li­metres and contain a maxi­mum of 15,000 neu­rons. Making them big­ger means pro­tec­ting them against oxy­gen depri­va­tion, to which neu­rons are very sen­si­tive. To do this, they need to be per­fu­sed, to ensure that each cell is connec­ted to a sup­ply of oxy­gen and nutrients, just as the blood­stream does in vivo.  Micro­flui­dics seems capable of pro­vi­ding this per­fu­sion. But this tech­no­lo­gy has not yet been trans­po­sed to cere­bral orga­noids. This tech­no­lo­gi­cal trans­fer will make it pos­sible to go from a culture the size of a fly’s brain to one com­pa­rable to that of a mouse.

Will these sys­tems replace our com­pu­ters ? Not in the near future. In 2019, a Japa­nese team suc­cee­ded in get­ting two cere­bral orga­noids to com­mu­ni­cate³. It is dif­fi­cult to pre­dict what will hap­pen next, given the large num­ber of teams wor­king on the sub­ject. Never­the­less, there is hope that the first OI sys­tems will be tools for neu­ros­cience research. Bio­lo­gi­cal com­pu­ters could help to explain how a brain manages to pro­cess incom­plete infor­ma­tion, for example.

They could also help to elu­ci­date the mecha­nisms of demen­tia, Asperger’s syn­drome, or other com­mon human brain condi­tions. It is cur­rent­ly very dif­fi­cult to have accep­table labo­ra­to­ry models for these issues. Ethi­cal consi­de­ra­tions, for example, natu­ral­ly for­bid stu­dying the effect of a mole­cule that dis­rupts memo­ry on humans… Cere­bral orga­noids offer an alter­na­tive for such research.

OI also raises ethi­cal ques­tions in its own right : Which sys­tems could be at risk of suf­fe­ring pain ? What pro­to­cols will be used to assess the intel­li­gence of a culture ? These issues need to be anti­ci­pa­ted, and OI spe­cia­lists have com­mit­ted to incor­po­ra­ting them into the deve­lop­ment of tech­ni­cal aspects by signing the Bal­ti­more Conven­tion in 2022⁴. The deve­lop­ment of the bio­lo­gi­cal com­pu­ter is still in its infancy.

Agnès Vernet