During my PhD I investigated how shape and motion information are processed by the
rat visual system, so as to establish how advanced is the representation of higher-order visual
information in this species and, ultimately, to understand to what extent rats can present a
valuable alternative to monkeys, as experimental models, in vision studies. Specifically, in my
thesis work, I have investigated:
1) The possible visual strategies underlying shape recognition.
2) The ability of rat visual cortical areas to represent motion and shape information.
My work contemplated two different, but complementary experimental approaches:
psychophysical measurements of the rat’s recognition ability and strategy, and in vivo
extracellular recordings in anaesthetized animals passively exposed to various (static and
moving) visual stimulation.
The first approach implied training the rats to an invariant object recognition task, i.e. to
tolerate different ranges of transformations in the object’s appearance, and the application of
an mage classification technique known as The Bubbles to reveal the visual strategy the
animals were able, under different conditions of stimulus discriminability, to adopt in order to
perform the task.
The second approach involved electrophysiological exploration of different visual areas
in the rat’s cortex, in order to investigate putative functional hierarchies (or streams of
processing) in the computation of motion and shape information.
Results show, on one hand, that rats are able, under conditions of highly stimulus
discriminability, to adopt a shape-based, view-invariant, multi-featural recognition strategy;
on the other hand, the functional properties of neurons recorded from different visual areas
suggest the presence of a putative shape-based, ventral-like stream of processing in the rat’s
visual cortex.
The general purpose of my work is and has been the unveiling the neural mechanisms
that make object recognition happen, with the goal of eventually 1) be able to relate my
findings on rats to those on more visually-advanced species, such as human and non-human
primates; and 2) collect enough biological data to support the artificial simulation of visual
recognition processes, which still presents an important scientific challenge
