ACTIVE SNIFFING · The Algorithmic and Neuronal Basis of Active Sensing in the Case of Odor-Trail Tracking
7РП — „Хора“ (Действия „Мария Кюри“)
- Период
- 2014-08-01 → 2017-11-30
- Финансиране от ЕС
- 252 290 €
- Участници
- 1
- Схема
- MC-IOF
Линиите свързват координатора с партньорите.
Накратко на български
Невронните механизми на активното усещане се изследват чрез проследяване на миризливи следи от мишки. Това помага да се разбере как животните вземат решения къде и кога да подушат, за да се ориентират в средата си.
Кратко обяснение, генерирано от езиков модел по текста на CORDIS. Оригиналът е по-долу.
Резултати накратко
The Algorithmic and Neuronal Basis of Active Sensing in the Case of Odor-Trail Tracking
Sensory perception is an active process. For instance, during normal vision an animal makes quick movements of the eye and fixates its gaze to a specific region of a visual scene. The visual system actively determines what parts of the visual environment it is going to sample. This common behavior, called active sensing, is not unique to the visual system, in fact many sensory systems operate in such a fashion and olfaction, the sense of smell, is no exception. Olfaction in rodents is an important model for active sensing. Rodents excel at the sense of smell, which is critical for finding mates and food, as well as for avoiding predators and toxins. Rodents smell by tracking either air-borne odors or surface-borne trails on the ground. During tracking, the animal decides when and where to sample space thereby updating an internal representation of its environment that ultimately guides its behavior. This sampling is achieved by ’sniffing’, which is defined as voluntary inhalation, and can happen at frequencies of up to four times the baseline respiration rate. This project investigated active sensing in the context of olfactory trail tracking in mice and its neuronal basis. To effectively tackle this question, we developed a novel high-throughput experimental system that allows the delivery of odor trails to a belt in real-time by using inkjet-printing technology. Subsequently we demonstrated that after a short habituation phase, mice will track the printed odor trails, if one also delivers rewards to the trail at random intervals. We have also established deep learning methods to accurately reconstruct the movements of the snout in 3D space during trail tracking from videography, which together with wireless sniff recordings allowed us to accurately measure where and when mice sampled the odor trail without obstructing the natural movement of mice. Ongoing efforts remain dedicated to analyze mouse behavior during trail tracking towards characterizing their trail tracking behavior.
Текст от CORDIS, на английски · Данни: CORDIS, © Европейски съюз
Цел на проекта
In its natural environment, any behaving animal is constantly bombarded with sensory information that must be parsed quickly by the animal to act appropriately. Yet, the nervous system is limited in the amount of information it can handle. Thus, to avoid sensory overload, specific aspects of these information streams have to be actively attended to by the animal. Such focusing is called active sensing and is a fundamental operational mode of many sensory systems. One ethologically important model for active sensing that has received increasing attention is odor-guided navigation in rodents. During odor-trail tracking, sniffing rates rise up to four times above the normal respiration frequency and each sniff selectively determines which part of the external world is sampled. Yet, an algorithmic understanding of their search strategy as well as its neuronal basis is lacking. My aims are to understand the neural representation of odors during this natural behavior and to reveal the neuronal basis of odor tracking behavior by following a multidisciplinary research program. I will develop an experimental setup that can flexibly deliver tightly controlled odor trails on a custom-build treadmill. By generating specifically tailored odor stimuli, I will be able to probe odor-guided navigation in real time and thus uncover the tracking algorithm. Moreover, I will record neuronal population activity from the early olfactory areas in mice that are following the trail. Such population activity in behaving animals can now be measured because of recent developments in miniature microscopes and genetically encoded calcium probes. I will show which aspects of the stimulus and behavioral variables are encoded by applying feature selection methods. Furthermore, by linking the behavioral decisions an animal takes during tracking to the recorded population response I will reveal which aspects of neuronal activity contribute to the tracking behavior in a mechanistic way.
Оригинален текст от CORDIS (на английски).
Участници
- EBERHARD KARLS UNIVERSITAET TUEBINGEN · TuebingenКоординаторГермания
Връзки
Данни: CORDIS, © Европейски съюз