kth.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Complete representation of action space and value in all dorsal striatal pathways
Department of Neuroscience, Karolinska Institutet, 171 77 Stockholm, Sweden.ORCID iD: 0000-0002-9173-7459
Department of Neuroscience, Karolinska Institutet, 171 77 Stockholm, Sweden.ORCID iD: 0000-0002-4754-4561
Department of Neuroscience, Karolinska Institutet, 171 77 Stockholm, Sweden.ORCID iD: 0000-0002-4578-2347
Department of Neuroscience, Karolinska Institutet, 171 77 Stockholm, Sweden.ORCID iD: 0000-0002-6122-9744
Show others and affiliations
2021 (English)In: Cell Reports, E-ISSN 2211-1247, Vol. 36, no 4, article id 109437Article in journal (Refereed) Published
Abstract [en]

The dorsal striatum plays a central role in the selection, execution, and evaluation of actions. An emerging model attributes action selection to the matrix and evaluation to the striosome compartment. Here, we use large-scale cell-type-specific calcium imaging to determine the activity of striatal projection neurons (SPNs) during motor and decision behaviors in the three major outputs of the dorsomedial striatum: Oprm1+ striosome versus D1+ direct and A2A+ indirect pathway SPNs. We find that Oprm1+ SPNs show complex tunings to simple movements and value-guided actions, which are conserved across many sessions in a single task but remap between contexts. During decision making, the SPN tuning profiles form a complete representation in which sequential SPN activity jointly encodes task progress and value. We propose that the three major output pathways in the dorsomedial striatum share a similarly complete representation of the entire action space, including task- and phase-specific signals of action value and choice.

Place, publisher, year, edition, pages
Elsevier BV , 2021. Vol. 36, no 4, article id 109437
National Category
Neurosciences
Identifiers
URN: urn:nbn:se:kth:diva-338934DOI: 10.1016/j.celrep.2021.109437ISI: 000678096100013PubMedID: 34320355Scopus ID: 2-s2.0-85111343842OAI: oai:DiVA.org:kth-338934DiVA, id: diva2:1808393
Note

QC 20231101

Available from: 2023-10-31 Created: 2023-10-31 Last updated: 2024-01-17Bibliographically approved
In thesis
1. On learning in mice and machines: continuous population codes in natural and artificial neural networks
Open this publication in new window or tab >>On learning in mice and machines: continuous population codes in natural and artificial neural networks
2023 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

Neural networks, whether artificial in a computer or natural in the brain, could represent information either using discrete symbols or continuous vector spaces. In this thesis, I explore how neural networks can represent continuous vector spaces, using both simulated neural networks and analysis of real neural population data recorded from mice. A special focus is on the networks of the basal ganglia circuit and on reinforcement learning, i.e., learning from rewards and punishments.

The thesis includes four scientific papers: two theoretical/computational (Papers I and IV) and two with analysis of real data (Papers II and III).

In Paper I, we explore methods for implementing continuous vector spaces in networks of spiking neurons using multidimensional attractors, and propose an explanation for why it is hard to escape the neural manifolds created by such attractors.

In Paper II, we analyze experimental data from dorsomedial striatum collected using 1-photon calcium imaging of transgenic mice with celltype-specific markers for the striatal direct, indirect and patch pathways, as the mice were gathering rewards in a 2-choice task. In line with extensive previous results, our data analysis revealed a number of neural signatures of reinforcement learning, but no apparent difference between the pathways.

In Paper III, we present a new software tool for tracking neurons across weeks of 1-photon calcium imaging, and employ it to follow patch-specific striatal projection neurons from the dorsomedial striatum across two weeks of daily recordings.

In Paper IV, we propose a model for how the nigrostriatal dopaminergic projection could, in a biologically plausible way, convey a vector-valued error gradient to the dorsal striatum, as required for backpropagation.

Based on the results of the papers and a review of existing literature, I argue that while the basal ganglia indeed make up a circuit for reinforcement learning as previously thought, this circuit represents reinforcement learning states, actions and policies using a continuous population code and not using discrete symbols.

Abstract [sv]

Neurala nätverk, så väl artificiella i en dator som naturliga i en hjärna, skulle kunna representera information antingen som diskreta symboler eller som steglösa vektorrymder. I denna avhandling utforskar jag hur neurala nätverk kan representera steglösa vektorrymder både genom att simulera neurala nätverk och genom att analysera uppmätt neural data från möss. Jag fokuserar i synnerhet på hjärnnätverken i de basala ganglierna och på förstärkningsinlärning, det vill säga inlärning baserad på belöning och bestraffning.

Avhandlingen innefattar fyra forskningsartiklar: två teoretiska med simuleringar (Artikel I och Artikel IV) och två med analys av uppmätt data (Artikel II och Artikel III).

I Artikel I utforskar vi metoder för att implementera steglösa vektorrymder i nätverk med spikande nervceller genom att skapa multidimensionella attraktorer, och lägger fram en förklaring för varför det är svårt att ta sig ur sådana attraktorer.

I Artikel II analyserar vi uppmätt hjärnaktivitet som spelats in från dorsomediala striatum med hjälp av kalciumavbildning (en-foton) i genmodifierade möss med markörer för den direkta, den indirekta respektive den striosoma nervbanan, medan mössen samlade belöningar i ett spel med två val. Likt omfattande tidigare forskning fann vi ett antal neurala signaturer för förstärkningsinlärning, men inga skillnader i signaturerna mellan de tre nervbanorna.

I Artikel III presenterar vi ett nytt mjukvaruverktyg för att följa nervceller genom veckolånga kalciuminspelningar och använder det specifikt till att följa striatala projektionsceller i den striosoma nervbanan under två veckors tid.

I Artikel IV presenterar vi en modell för hur den nigrostriatala dopaminerga nervbanan skulle kunna förmedla en vektorvärd felgradient till dorsomediala striatum på ett biologiskt trovärdigt sätt. En sådan behövs för backpropagering.

Utifrån resultaten i artiklarna och en litteraturöversikt drar jag slutsatsen att basala ganglierna implementerar en förstärkningsinlärningsalgoritm i enlighet med tidigare forskning, men att representationen av tillståndsrymd, handlingar, och policyer byggs upp av en kontinuerlig populationskod och inte av diskreta symboler.

Place, publisher, year, edition, pages
Stockholm: Karolinska Institutet, 2023. p. 136
National Category
Neurosciences Bioinformatics (Computational Biology)
Research subject
Biotechnology; Applied and Computational Mathematics
Identifiers
urn:nbn:se:kth:diva-338936 (URN)978-91-8017-152-6 (ISBN)
Public defence
2023-11-24, E1, Lindstedtsvägen 3, Stockholm, 13:00 (English)
Opponent
Supervisors
Note

Thesis presented for joint PhD degree between KI and KTH, with KI as home university.

QC 20231101

Available from: 2023-11-01 Created: 2023-10-31 Last updated: 2023-11-20Bibliographically approved

Open Access in DiVA

fulltext(46873 kB)115 downloads
File information
File name FULLTEXT01.pdfFile size 46873 kBChecksum SHA-512
24c7744a0f6f365aace50d41803c14c2de441961e9cb8474638bdb82f8350d98a2acabb2e438c2c770212719e35649b98472f468256820aa0627ea12391f54d1
Type fulltextMimetype application/pdf

Other links

Publisher's full textPubMedScopus

Authority records

Wärnberg, Emil

Search in DiVA

By author/editor
Weglage, MoritzWärnberg, EmilLazaridis, IakovosCalvigioni, DanielaMeletis, Konstantinos
In the same journal
Cell Reports
Neurosciences

Search outside of DiVA

GoogleGoogle Scholar
Total: 115 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

doi
pubmed
urn-nbn

Altmetric score

doi
pubmed
urn-nbn
Total: 259 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf