Habit learning and brain–machine interfaces (BMI): a tribute to Valentino Braitenberg’s “Vehicles”

Niels Birbaumer, Friedhelm C. Hummel

Research output: Contribution to journalArticlepeer-review


Brain–Machine Interfaces (BMI) allow manipulation of external devices and computers directly with brain activity without involvement of overt motor actions. The neurophysiological principles of such robotic brain devices and BMIs follow Hebbian learning rules as described and realized by Valentino Braitenberg in his book “Vehicles,” in the concept of a “thought pump” residing in subcortical basal ganglia structures. We describe here the application of BMIs for brain communication in totally locked-in patients and argue that the thought pump may extinguish—at least partially—in those people because of extinction of instrumentally learned cognitive responses and brain responses. We show that Pavlovian semantic conditioning may allow brain communication even in the completely paralyzed who does not show response-effect contingencies. Principles of skill learning and habit acquisition as formulated by Braitenberg are the building blocks of BMIs and neuroprostheses.

Original languageEnglish
Pages (from-to)595-601
Number of pages7
JournalBiological Cybernetics
Issue number5
Publication statusPublished - 2014


  • Instrumental and classical conditioning
  • Neuroprosthetics
  • Skill learning

ASJC Scopus subject areas

  • Biotechnology
  • Computer Science(all)
  • Medicine(all)


Dive into the research topics of 'Habit learning and brain–machine interfaces (BMI): a tribute to Valentino Braitenberg’s “Vehicles”'. Together they form a unique fingerprint.

Cite this