Biology Reference
In-Depth Information
The fact that spatial patterns, which imply continuous representations of space, are
represented in the hippocampus has led to the application of continuous attractor models to
help understand hippocampal function (Fig 6). A class of network that can maintain the firing
of its neurons to represent any location along a continuous physical dimension such as spatial
position and head direction is a ''Continuous Attractor'' neural network (Stringer
et al.
, 2004;
Rolls & Kesner, 2006). It uses excitatory recurrent collateral connections between the
neurons, as in the case with CA3, to reflect the distance between the neurons that represent
allocentric spatial configuration.
The main function of an autoassociative network is to produce the correct firing of all
cells that encode a memory when presented with only a partial or degraded form of that
memory. Abstract theoretical models of sequence recall suggested that accurate sequence
recall could be achieved by having autoassociative processes interact with heteroassociative
processes (Kleinfeld, 1986; Sompolinsky & Kanter, 1986), and this concept was adapted to
the specific circuitry of the hippocampus (Lisman, 1999; Lisman & Otmakhova, 2001). A
possible CA3 recurrent network function is to store “heteroassociations” that link the cells
encoding sequential memory items. The main function of a heteroassociative network is to
recall the subsequent memory items in particular order when presented with a memory cue
(Lisman & Otmakhova, 2001). In summary, it is proposed that the reciprocally interacting
heteroassociative and autoassociative networks produce more accurate in learning and
recalling sequences (Lisman, 1999).
6.
P
LACE
F
IELDS AND
S
EQUENCE
M
EMORY
Hippocampal circuity could store and recall memory sequences and a major line of
evidence for sequence recall is the “phase precession” of hippocampal place cells. As the rat
enters the receptive field of the neuron, the spikes occur on the peak of the theta cycle and
may precede a full period as the rat passes through the entire receptive field of the cell
(Dragoi & Buzsáki, 2006). Sequential activation of hippocampal place cells on a track can be
represented by unique sets of cell assemblies (Fig 7), which are bound together by synaptic
interactions into an episode (Jensen & Lisman, 1996; Tsodyks
et al.
, 1996). Such
organization implies temporally-coordinated activity within and between anatomically
distributed groups of sequential cell assemblies. Acting as an attractor dynamic system
hippocampal formation induces phase precession of spikes within the theta cycle (Jensen &
Lisman, 1996; Tsodyks
et al.
, 1996; Samsonovich & McNaughton, 1997; Wallenstein &
Hasselmo, 1997; Wills
et al.
, 2005). Spike-phase variability of the place cells is temporally-
correlated as the timing of neuronal action potentials depends on the activity of the
synaptically-connected cell assemblies in which individual cells are embedded. The great
majority of intrahippocampal synapses is established by the collateral system of CA3 neurons
(Amaral & Witter, 1989; Li
et al.
, 1994), and it has been hypothesized that distances between
place fields are encoded in the synaptic strengths between CA3-CA3 and CA3-CA1 neuron
pairs (Muller
et al.
, 1996). Temporal encoding of spatial information, therefore, can be
explained by the experience-dependent modification of synaptic strengths in these regions. It
is hypothesized that the sequences are stored in the autoassociative CA3 recurrent and CA3-
CA1 collateral systems (Jensen & Lisman, 1996; Muller
et al.
, 1996; Tsodyks, 1999; Dragoi
