研究 - 麻省理工学院人工智能研究专业

研究

Modeling the inﬂuence of the hippocampal
memory system on the oculomotor system

Jennifer D. Ryan

1
John Grifﬁths

1,2∗

, Kelly Shen
3
, Gleb Bezgin

1∗

1
, Arber Kacollja

, Heather Tian

, and Anthony R. McIntosh

1,2

1Rotman Research Institute, Baycrest, 多伦多, 安大略省, 加拿大
2心理学系, 多伦多大学, 多伦多, 安大略省, 加拿大
Montreal Neurological Institute, 麦吉尔大学, 蒙特利尔, Quebec, 加拿大

关键词: Hippocampus, Eye movements, Functional Dynamics, 记忆, Oculomotor guidance

开放访问

杂志

抽象的

Visual exploration is related to activity in the hippocampus (HC) and/or extended medial
temporal lobe system (MTL), is inﬂuenced by stored memories, and is altered in amnesic
案例. An extensive set of polysynaptic connections exists both within and between the HC
and oculomotor systems such that investigating how HC responses ultimately inﬂuence
neural activity in the oculomotor system, and the timing by which such neural modulation
could occur, is not trivial. We leveraged TheVirtualBrain, a software platform for large-scale
network simulations, to model the functional dynamics that govern the interactions between
the two systems in the macaque cortex. Evoked responses following the stimulation of the
MTL and some, but not all, subﬁelds of the HC resulted in observable responses in
oculomotor regions, including the frontal eye ﬁelds, within the time of a gaze ﬁxation.
Modeled lesions to some MTL regions slowed the dissipation of HC signal to oculomotor
地区, whereas HC lesions generally did not affect the rapid MTL activity propagation to
oculomotor regions. These ﬁndings provide a framework for investigating how information
represented by the HC/MTL may inﬂuence the oculomotor system during a ﬁxation and
predict how HC lesions may affect visual exploration.

作者总结

No major account of oculomotor (eye movement) guidance considers the inﬂuence of the
hippocampus (HC) and broader medial temporal lobe (MTL) 系统, yet it is clear that
information is exchanged between the two systems. Prior experience inﬂuences current
viewing, and cases of amnesia due to compromised HC/MTL function show speciﬁc
alterations in viewing behavior. By modeling large-scale network dynamics, we show that
stimulation of subregions of the HC, and of the MTL, rapidly results in observable responses
in oculomotor control regions, and that HC/MTL lesions alter signal propagation. 这些
ﬁndings suggest that information from memory may readily guide visual exploration, and call
for a reconsideration of the neural circuitry involved in oculomotor guidance.

介绍

Memory inﬂuences ongoing active exploration of the visual environment (Hannula et al.,
2010). 例如, more viewing is directed to novel versus previously viewed items (费根,
1970; Fantz, 1964), and more viewing is directed to areas that have been altered from a
prior viewing (Ryan, Althoff, Whitlow, & 科恩, 2000; 史密斯, 霍普金斯, & Squire, 2006). A

引文: Ryan, J. D ., 沉, K., Kacollja,
A。, Tian, H。, Grifﬁths, J。, Bezgin, G。, &
McIntosh, A. 右. (2020). 建模
inﬂuence of the hippocampal memory
system on the oculomotor system.
网络神经科学, 4(1), 217–233.
https://doi.org/10.1162/netn_a_00120

DOI:
https://doi.org/10.1162/netn_a_00120

支持信息:
https://doi.org/10.1162/netn_a_00120

已收到: 31 七月 2019
公认: 4 十二月 2019

利益争夺: 作者有
声明不存在竞争利益
存在.

Corresponding Authors:
Jennifer D. Ryan
jryan@research.baycrest.org
Kelly Shen
kshen@research.baycrest.org

处理编辑器:
Michael Cole

版权: © 2019
麻省理工学院
在知识共享下发布
归因 4.0 国际的
(抄送 4.0) 执照

麻省理工学院出版社

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
n
e
n
A
r
t
我
C
e
–
p
d

我

F
/

4
1
2
1
7
1
8
6
6
7
5
4
n
e
n
_
A
_
0
0
1
2
0
p
d

乙
y
G
你
e
s
t

哦
n
0
9
S
e
p
e
米
乙
e
r
2
0
2
3

Simulated dynamics of hippocampal responses in oculomotor regions

Hippocampus:
A region within the medial temporal
lobe of the brain that has a critical
role in the formation of memory
representations that consist of the
relations among distinct elements.

Medial temporal lobe:
A collection of neural regions that
includes the hippocampus, subicular
复杂的, and surrounding entorhinal
cortex, parahippocampal cortex, 和
perirhinal cortex, which collectively
support cognitive functions such as
洞察力, 导航, and memory.

Visual exploration:
The manner by which the eyes move
to explore the visual world. Entails
the active process of alternating
ballistic movements of the eyes
(saccades)and pauses by the eyes in
order to inspect regions in space
(gaze ﬁxations).

Oculomotor system:
A distributed network of neural
regions that provide cognitive and
motor control over the movement of
the eyes.

number of studies have implicated a network of subregions within the hippocampus (HC)
and/or broader medial temporal lobe (MTL) responsible for the inﬂuence of memory on viewing
行为. Amnesic cases who have severe memory impairments due to compromised function
of the HC and/or MTL show changes in their viewing behavior compared with neurologically
intact cases (Chau, 墨菲, Rosenbaum, Ryan, & Hoffman, 2011; Hannula, Ryan, Tranel, &
科恩, 2007; Olsen et al., 2015; Ryan et al., 2000; Warren, Duff, Tranel, & 科恩, 2010).
Similar ﬁndings have been observed in older adults who have suspected HC/MTL compromise
(Ryan, Leung, Turk-Browne, & Hasher, 2007), and certain viewing patterns have been shown
to track with entorhinal cortex (ERC) volumes (Yeung et al., 2017). Visual exploration predicts
HC activity during encoding (刘, 沉, Olsen, & Ryan, 2017), 和, 反过来, HC/MTL ac-
tivity predicts ongoing visual exploration that is indicative of memory retrieval (Hannula &
Ranganath, 2009; Ryals, 王, Polnaszek, & Voss, 2015). The relationship between visual
sampling and HC activity is weakened in aging, presumably because of decline in HC struc-
ture or function (刘, 沉, Olsen, & Ryan, 2018). Such evidence collectively demonstrates that
HC/MTL function is related to oculomotor behavior. The indirect implication of these studies
is that the HC must inﬂuence neural activity in the oculomotor system.

Studies in nonhuman primates have shown that HC/MTL activity is linked to oculomotor
行为. The activity of grid cells in the ERC is tied to eye position (Killian, Jutras, & Buffalo,
2012), while HC/MTL activity is modulated by saccades (Sobotka, Nowicka, & Ringo, 1997)
and ﬁxations (Hoffman et al., 2013; Leonard et al., 2015). How HC/MTL activity traverses the
brain to inﬂuence the oculomotor system has not been shown to date. The oculomotor system
is itself a highly recurrent and distributed network (Parr & 弗里斯顿, 2017) composed of corti-
cal and subcortical regions responsible for the execution of a saccade (例如, frontal eye ﬁeld,
FEF; superior colliculus, SC) as well as regions that exert cognitive control over where the eyes
should go (例如, dorsolateral prefrontal cortex, dlPFC; anterior cingulate cortex, ACC; lateral
intraparietal area, area LIP) (Bisley & Mirpour, 2019; Johnston & Everling, 2008). Prior work
has speculated as to which regions of the brain may be important for bridging the memory and
oculomotor systems (例如, Meister & Buffalo, 2016; Micic, Ehrlichman, & 陈, 2010), 但
these discussions were limited to regions examined in isolation. There are no known direct
connections between hippocampal subﬁelds and the oculomotor system. 然而, by examining
whole-cortex connectivity, we have shown that there is an extensive set of polysynaptic path-
ways spanning extrastriate, 后顶叶, and prefrontal regions that may mediate the ex-
change of information between the oculomotor and memory systems (沉, Bezgin, Selvam,
McIntosh, & Ryan, 2016). Given the vast anatomical connectivity within and between the
memory and oculomotor systems, trying to discern the functional network involved in bridg-
ing them is not a trivial problem to tackle. Speciﬁcally, the large and complex contribution
of recurrent anatomical connections to the functional dynamics of large-scale brain networks
must be considered (Spiegler, 汉森, Bernard, McIntosh, & Jirsa, 2016). One crucial ques-
tion concerning such functional dynamics is whether HC/MTL activity is able to inﬂuence the
activity related to the preparation of a saccade. To impact ongoing visual exploration, HC/MTL
activity would likely need to resolve in the oculomotor system within the time of an average
duration of a gaze ﬁxation (∼250–400 ms) (Henderson, Nuthmann, & 卢克, 2013).

To examine the extent to which HC/MTL activity could inﬂuence the oculomotor system,
we leveraged a computational modeling and neuroinformatics platform, TheVirtualBrain, 和
simulated the functional dynamics of a whole-cortex directed macaque network when stim-
ulation is applied to HC and MTL nodes of interest. Critically, we examined whether and
when evoked activity culminated in responses in key regions within the oculomotor system.
最后, we observed the extent to which the propagation and timing of such activity was altered

网络神经科学

218

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
n
e
n
A
r
t
我
C
e
–
p
d

我

F
/

4
1
2
1
7
1
8
6
6
7
5
4
n
e
n
_
A
_
0
0
1
2
0
p
d

乙
y
G
你
e
s
t

哦
n
0
9
S
e
p
e
米
乙
e
r
2
0
2
3

Simulated dynamics of hippocampal responses in oculomotor regions

Amnesia:
A severe memory disorder that arises
because of lesion or dysfunction of
the hippocampus and/or surrounding
medial temporal lobe. Individuals
with amnesia may be unable to
retrieve previously established
回忆 (retrograde amnesia) 和
form new memories (anterograde
amnesia).

following lesions to one or more HC/MTL regions in order to understand the neural dynam-
ics that may underlie altered visual exploration in cases of HC/MTL dysfunction, 比如在
amnesia or aging.

结果

We modeled the inﬂuence of HC/MTL activity on the oculomotor system using a connectome-
based approach using TheVirtualBrain (see Methods for details). Following Spiegler and col-
联赛 (Spiegler et al., 2016), we assigned a neural mass model to each node and set each to
operate near criticality, which is considered to be the point at which information processing
capacity is maximal (Deco et al., 2014; 戈什, Rho, McIntosh, Kötter, & Jirsa, 2008). Nodes
were then connected together as deﬁned by a weighted and directed macaque structural con-
nectivity matrix and the distance between them deﬁned by a tract lengths matrix. As we were
interested in examining signal propagation from the memory system to the oculomotor system
while taking into account the extensive recurrent connectivity between them, we chose to use
the macaque connectome because of the available information from tracer data on the direc-
tionality of ﬁber tracts. Without stimulation, this network exhibits no activity. 然而, 和
stimulation, activity dissipates throughout the network according to the spatiotemporal con-
straints imposed by the connectivity weights and distances. Despite having no spontaneous
活动, this model has been shown to exhibit the emergent properties of spontaneous activity
(Spiegler et al., 2016). 那是, with stimulation, the model produces a diverse set of resting-state
networks that are typically detected from spontaneous activity in empirical studies. Cortical
network dynamics were set via additional parameter tuning such that stimulating V1 resulted
in biologically plausible timing of evoked responses in downstream visual cortical regions.
最后, we systematically stimulated HC subﬁelds and MTL regions of interest and detected
evoked responses across the rest of the network.

HC Subregion Stimulation

Stimulation of HC subﬁelds and MTL regions of interest evoked widespread activation across
网络, similar to previous surface-based model simulations (Spiegler et al., 2016).
数字 1 shows an example of activity dissipation following CA1 stimulation. Evoked responses
were ﬁrst detected in other HC subﬁelds and MTL regions but then spread to prefrontal and
extrastriate cortices, and later to posterior parietal cortex. The full list of activation times for
每一个 77 nodes can be found in Supporting Information Table S2. 然而, in all sub-
sequent analyses, we present only the results pertaining to our nodes of interest, identiﬁed as
those along the shortest paths between HC/MTL and oculomotor regions (Shen et al., 2016) 或者
those that have been speciﬁcally suggested in the literature to be potentially relevant (Meister
& Buffalo, 2016).

Stimulation of HC subregions CA1, subiculum (S), presubiculum (PrS), and parasubiculum
(PaS) resulted in observable responses in almost all of the cortical nodes of interest, and within
地区 46, 24, and FEF, of the oculomotor system (for CA1 example; see Figure 2A). 之内
our oculomotor regions of interest, activity was ﬁrst observed in area 46, 其次是 24, 和
FEF, regardless of HC stimulation site. Stimulation of the PrS resulted in the fastest observable
responses in these oculomotor areas (在下面 70 多发性硬化症; Figure 3A). Stimulation of CA1 resulted in
rapid activity that culminated in oculomotor regions in under 220 多发性硬化症 (Figure 3B). Stimulation
of either the S or the PaS resolved into area 46 activity by 200 多发性硬化症, into area 24 经过 250 多发性硬化症, 和
ﬁnally into FEF by 500 多发性硬化症 (Figure 3C and 3D). Responses in area LIP occurred substantially
later than the other oculomotor areas and even later than all other cortical nodes for CA1 and

网络神经科学

219

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
n
e
n
A
r
t
我
C
e
–
p
d

我

F
/

4
1
2
1
7
1
8
6
6
7
5
4
n
e
n
_
A
_
0
0
1
2
0
p
d

乙
y
G
你
e
s
t

哦
n
0
9
S
e
p
e
米
乙
e
r
2
0
2
3

Simulated dynamics of hippocampal responses in oculomotor regions

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
n
e
n
A
r
t
我
C
e
–
p
d

我

F
/

4
1
2
1
7
1
8
6
6
7
5
4
n
e
n
_
A
_
0
0
1
2
0
p
d

乙
y
G
你
e
s
t

哦
n
0
9
S
e
p
e
米
乙
e
r
2
0
2
3

数字 1. Dissipation of activity over time across the cortex following simulated stimulation of
CA1. Average above-threshold simulated activation (arbitrary units) for each node for a 10-ms epoch
following each time point is plotted on the macaque cortical surface. Activations were log scaled
for the purposes of visualization. 从左到右: ventral, dorsal, medial, and lateral views of the
macaque cortical surface.

PrS stimulations (>440 ms; 桌子 1). No evoked response was detected in area LIP following
stimulation of S or PaS. Responses were not observed in the majority of the predeﬁned cortical
hubs following CA3 stimulation, and activity did not culminate in observable responses in the
oculomotor areas (Figure 3B). 见表 1 for activation times for all nodes of interest.

MTL Stimulation

Stimulation of any of the broader regions within the MTL (entorhinal cortex, ERC; perirhinal
cortex, 35, 36; parahippocampal cortex, TF, TH) resulted in observable responses within ocu-
lomotor areas 46, 24, and FEF well under 100 多发性硬化症, faster than the responses observed from
HC subﬁeld stimulation. Of the MTL regions, stimulation of area 35/36 resulted in the earliest
responses in areas 46, 24, and FEF (之内 25 多发性硬化症). Although evoked responses in area LIP
occurred in under ∼250 ms for all but ERC stimulations, area LIP again exhibited the most
delayed response across all nodes of interest following MTL stimulations. 见表 1 for acti-
vation times for all nodes of interest.

网络神经科学

220

Simulated dynamics of hippocampal responses in oculomotor regions

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
n
e
n
A
r
t
我
C
e
–
p
d

我

F
/

4
1
2
1
7
1
8
6
6
7
5
4
n
e
n
_
A
_
0
0
1
2
0
p
d

乙
y
G
你
e
s
t

哦
n
0
9
S
e
p
e
米
乙
e
r
2
0
2
3

(A) Simulated stimulation of the CA1 (red circle) resulted in observable responses (蓝色的
数字 2.
circles) in multiple hippocampus (HC)/medial temporal lobe (MTL) 节点, intermediary nodes, 和
in regions governing oculomotor control, including the frontal eye ﬁelds (FEFs). (乙) Simulated stim-
ulation of the CA3 (red circle) resulted in observable responses (blue circles) limited to HC/MTL
节点. Very few responses were observed in cortical areas and none were observed in oculomotor
地区. Size and shade of the circles scale with elapsed time prior to an observed response. Gray
lines denote direct structural connections between nodes. For visualization purposes, only regions
that contribute to the shortest paths from HC/MTL to oculomotor nodes are shown. 连接
between intermediary nodes (middle column) are not shown. Connections that are unidirectional
and away from oculomotor areas (IE。, to HC/MTL) are indicated by dashed lines.

Cortical Responses

HC and MTL region stimulation (except for CA3) resulted in signal propagation across all of
our preidentiﬁed cortical regions of interest. When CA3 was stimulated, cortical responses
were only observed in areas V2 and 25; no other signal was observed. 尤其, responses in
地区 5 and 7a were generally observed following activity from oculomotor regions, 包括-
ing FEF, suggestive of a possible feedback response. The exception is S stimulation, 其中
responses in area 5 preceded responses in oculomotor regions by ∼100 ms. Responses in V4
also followed oculomotor responses, except in cases of CA1, TF, and TH stimulation. 辛-
伊拉利, responses in area 23 followed oculomotor responses, except in cases of PrS and TH
stimulation. 见表 1 for activation times for all nodes of interest.

Lesion Models: HC Subregions

Some models of HC and MTL lesions showed an appreciable effect on activation times while
others did not. Only the results for lesions that affected any activation time by at least ±10 ms
are shown. Lesion of CA3 changed neither the pattern nor the timing of observable responses
following stimulation of each of the other HC/MTL regions (data not shown). Lesion of CA1

网络神经科学

221

Simulated dynamics of hippocampal responses in oculomotor regions

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
n
e
n
A
r
t
我
C
e
–
p
d

我

F
/

4
1
2
1
7
1
8
6
6
7
5
4
n
e
n
_
A
_
0
0
1
2
0
p
d

乙
y
G
你
e
s
t

哦
n
0
9
S
e
p
e
米
乙
e
r
2
0
2
3

数字 3. Simulated response proﬁles (envelope of region time series) of oculomotor areas fol-
lowing stimulation of PrS (A), CA1 (乙), S (C), and PaS (D). Activation is given in arbitrary units (a.u.).
The onsets of the responses for each oculomotor area indicated by arrows. Area LIP did not exhibit
a response that exceeded its baseline threshold following S and PaS stimulation.

网络神经科学

222

Simulated dynamics of hippocampal responses in oculomotor regions

桌子 1. Simulated activation times (多发性硬化症) following stimulation of hippocampal subﬁelds and medial
temporal lobe regions.

Stimulated node

e
d
哦
n

n
哦
我
t
A
v
r
e
s
乙
氧

CA3
CA1
S
PrS
PaS
ERC
35
36
TF
TH
5
10
11
12
13
14
23
25
32
Ig
Pro
7A
FST
MIP
PO
MSTd
公吨
V4
副总裁
V3
V2
24
46
FEF
LIP

CA3
0
137
115
84
350
169
567
317
137
7
N/A
N/A
N/A
N/A
N/A
N/A
N/A
435
N/A
N/A
N/A
N/A
186
N/A
656
N/A
N/A
N/A
753
N/A
215
N/A
N/A
N/A
N/A

CA1
48
0
17
23
17
40
6
1
0
0
322
59
12
96
10
9
165
6
38
11
52
381
47
237
208
286
260
76
55
260
56
107
84
217
535

S
31
0
0
119
96
0
21
14
20
29
50
24
68
186
18
27
384
19
63
63
96
202
147
523
593
664
857
353
266
841
274
250
135
452
N/A

PrS
37
0
26
0
28
4
49
10
0
0
92
83
74
93
62
57
12
35
88
75
12
336
52
215
190
287
272
103
86
257
48
55
15
68
445

PaS
27
46
37
25
0
0
50
6
44
0
555
106
166
285
132
76
231
57
89
189
48
795
79
296
248
494
483
244
259
409
66
238
187
500
N/A

ERC
0
0
0
6
50
0
0
0
11
20
249
12
10
15
8
9
43
6
10
11
3
256
6
405
355
33
189
280
223
398
139
24
12
79
561

35
21
8
16
13
15
0
0
0
12
29
129
10
7
9
7
9
37
11
13
4
1
53
1
113
99
23
107
147
98
154
16
22
9
19
233

36
20
2
10
11
0
0
0
0
6
18
141
12
8
10
11
11
52
8
14
6
4
52
3
126
107
23
23
142
101
163
17
25
11
15
230

TF
48
0
0
0
0
7
9
6
0
0
69
81
38
24
19
13
24
9
14
21
29
103
0
70
60
94
36
5
2
68
6
22
17
34
176

TH
1
0
0
0
32
13
19
14
0
0
64
20
48
34
19
20
13
12
17
13
20
59
6
48
39
53
83
15
41
68
0
16
23
71
144

笔记. Only nodes of interest (HC/MTL regions, oculomotor regions, and regions that are involved
in the shortest paths between HC/MTL and oculomotor nodes) are shown. S= subiculum; PrS =
presubiculum; PaS = parasubiculum; ERC = entorhinal cortex; 35/36 = perirhinal cortex; TF/TH =
parahippocampal cortex; 0 = stimulation onset; N/A = no response observed. For a comprehensive
set of activation times for all nodes in the network, see Supporting Information Table S2.

resulted in a lack of signal to V2, V4, 区域 23, and slowing of signal from the subicular com-
plex to various regions, including oculomotor regions FEF and area 24 (Figure 4A; 配套
Information Table S3). Lesion of CA1 also led to small increases in the speed of signal follow-
ing CA3 stimulation to the subicular complex, and from MTL regions to TF/TH, and to other
regions within the subicular complex (all less than 10 多发性硬化症).

网络神经科学

223

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
n
e
n
A
r
t
我
C
e
–
p
d

我

F
/

4
1
2
1
7
1
8
6
6
7
5
4
n
e
n
_
A
_
0
0
1
2
0
p
d

乙
y
G
你
e
s
t

哦
n
0
9
S
e
p
e
米
乙
e
r
2
0
2
3

Simulated dynamics of hippocampal responses in oculomotor regions

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
n
e
n
A
r
t
我
C
e
–
p
d

我

F
/

4
1
2
1
7
1
8
6
6
7
5
4
n
e
n
_
A
_
0
0
1
2
0
p
d

乙
y
G
你
e
s
t

哦
n
0
9
S
e
p
e
米
乙
e
r
2
0
2
3

数字 4. Changes in simulated activation times following HC lesions. Subicular stimulation fol-
lowing CA1 (A) and ERC (乙) 病变. Only nodes of interest are presented on the brain surface plots.
Activation time differences were computed by subtracting the prelesion activation times from the
postlesion ones. Absence of response following a lesion indicated in gray. 从左到右: ventral,
dorsal, medial, and lateral views of the macaque cortical surface.

Lesions to either the S or PaS produced little change to either the pattern or timing of re-
sponses following stimulation of the other HC/MTL regions (data not shown). Lesions to the
PrS produced moderate changes (<20 ms) in timing: there was some slowing of activity prop- agation from PaS to cortical regions, including oculomotor and speeding of signal propagation within the HC subﬁelds TF>140 多发性硬化症)
(Supporting Information Table S5).

Lesion Models: MTL Regions

Lesion of the ERC resulted in considerable slowing of observable signal in areas 24, 46, 和
FEF (30–340 ms) following S (Figure 3B) or PaS stimulation (Supporting Information Table S6).
TF and/or TH lesions resulted in slowing (10–400 ms) of signal following CA1, S, and PaS
stimulation to one or more of areas 24, 46, and FEF, and a lack of response in FEF following
PaS stimulation (only the combined TF/TH is shown; Supporting Information Table S7). Area
35 和/或 36 lesions also resulted in slowing (10–90 ms) of signal following CA1, S, and PaS
stimulation to one or more of areas 24, 46, and FEF, although not as severe as the slowing

网络神经科学

224

Simulated dynamics of hippocampal responses in oculomotor regions

observed following TF/TH lesions (only the combined 35/36 lesion is shown; 配套
Information Table S8).

Other Cortical Lesions

In our original stimulations, signals in regions 5, 7A, 23, and V4 were predominantly observed
following observable responses in oculomotor areas 24, 46, and FEF, suggesting these cortical
areas are receiving feedback signals rather than primarily serving as hubs to transfer signal
from the HC/MTL to the oculomotor regions. To explore this in more depth, we simulated a
combined lesion of 5/7a/23/V4 and examined signal propagation. Following this combined
cortical lesion, stimulation of each of the HC/MTL regions (except for CA3) continued to re-
sult in observable signal in areas 24, 46, and FEF. 有趣的是, while area LIP exhibited the
slowest overall responses in the intact simulations, the combined cortical lesion led to sub-
stantial speeding of signal to LIP from HC (200–270 ms faster) and MTL (80–300 ms faster) 关于-
gions owing to speeded responses in intermediary visual cortical and parietal areas (配套
Information Table S9).

讨论

A preponderance of evidence has demonstrated a correlation between HC/MTL neural activity
and oculomotor behavior (Hannula et al., 2010; Killian, Potter, & Buffalo, 2015; 刘等人。,
2017), but research had not shown whether HC/MTL activity can reach the oculomotor system
in time to inﬂuence the preparation of a saccade. The HC is well connected anatomically to
the oculomotor system through a set of polysynaptic pathways that span MTL, frontal, 顶叶,
and visual cortices (Shen et al., 2016), but the existence of anatomical connections does not
provide conclusive evidence of the functional relevance of speciﬁc pathways. By considering
the functional dynamics and recurrent interactions of the large-scale network involved in the
HC/MTL guidance of eye movements, we show that propagation of evoked HC/MTL neural
activity results in neural activity observable in areas 24 (ACC), 46 (dlPFC), and FEF, 哪个
are important for the cognitive and motoric control of eye movements, 分别 (Johnston
& Everling, 2008). Critically, the culmination of neural signal in these oculomotor regions
occurred within the time of a typical gaze ﬁxation (∼250–400 ms; Buswell, 1935; Henderson
等人。, 2013): 之内 200 ms following HC subﬁeld stimulation (except for CA3), and within
100 ms following stimulation of each MTL region. Our ﬁndings suggest that the underlying
neural dynamics of the memory and oculomotor systems allow for representations mediated
by the HC/MTL to guide visual exploration—what is foveated and when—on a moment-to-
moment basis.

The lack of responses in the FEF following CA3 stimulation is not surprising, given that there
are no known direct connections, and fewer polysynaptic pathways, between the CA3 and the
oculomotor regions investigated here (Shen et al., 2016). These functional and anatomical dif-
ferences align well with the purported representational functions of CA3 versus CA1. Foveated
information may be bound into detailed memory representations via the autoassociative net-
work of the CA3 (pattern separation; Norman & O’Reilly, 2003; Yassa & Stark, 2011), 然而
CA1 would enable the comparison of stored information to the external visual world ( pattern
completion; Rolls, 2013; Yassa & Stark, 2011).

Stimulation of the subiculum and parasubiculum resulted in relatively slower responses
observed in each of the oculomotor regions, whereas stimulation of presubiculum resulted in
rapid responses observed in the oculomotor regions. The subiculum and parasubiculum may
largely provide information that supports the grid cell mapping of the ERC (Boccara et al.,

网络神经科学

225

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
n
e
n
A
r
t
我
C
e
–
p
d

我

F
/

4
1
2
1
7
1
8
6
6
7
5
4
n
e
n
_
A
_
0
0
1
2
0
p
d

乙
y
G
你
e
s
t

哦
n
0
9
S
e
p
e
米
乙
e
r
2
0
2
3

Simulated dynamics of hippocampal responses in oculomotor regions

2010; Peyrache, Schieferstein, & Buzsáki, 2017; Tang et al., 2016). These regions may then
function as a “pointer” by providing online information of an individual’s location in space
(Tang et al., 2016). This slowly changing spatial layout may not then require a rapid inﬂuence
on the oculomotor system, 但反而, may allow for the presubiculum, which has cells that are
responsive to head direction (罗伯逊, Rolls, Georges-François, & Panzeri, 1999) to precisely
locate and foveate visual objects. These functional distinctions are speculative and remain to
be tested.

Stimulation of each of the MTL cortices resulted in observable responses in each of areas
24, 46, and FEF that were faster than any of the responses observed following HC subregion
stimulation. The MTL cortices are intermediary nodes that may permit the relatively rapid trans-
fer of information from HC to the oculomotor system. The unique representational content
supported by each region may inﬂuence ongoing visual exploration in a top-down manner.
The PRC provides lasting information regarding the features of objects (Erez, Cusack, 肯德尔,
& Barense, 2016; Graham, Barense, & 李, 2010), the PHC provides information regarding
the broader spatial environment (Alvarado & Bachevalier, 2005; Eichenbaum, Yonelinas, &
Ranganath, 2007; Sato & Nakamura, 2006), and the ERC may provide information regarding
the relative spatial arrangements of features within (Yeung et al., 2017), and among, 物体
within the environment (Buckmaster, 2004; Yeung et al., 2019). Signal from the MTL may be
used to accurately, and rapidly, prioritize gaze ﬁxations to areas of interest.

HC subﬁeld lesions only minimally altered the timing of activity from MTL to oculomotor
地区; the relatively rapid propagation of signal from MTL to FEF (<100 ms) was preserved. Lesions to MTL regions resulted in slowing of signal from some HC subﬁelds oculomotor regions. This pattern results suggests that different patterns visual exploration (i.e., rate, area) may occur cases>3.0.一氧化碳;2-H

Rolls, 乙. 时间. (2013). The mechanisms for pattern completion and pat-
tern separation in the hippocampus. Frontiers in Systems Neuro-
科学. https://doi.org/10.3389/fnsys.2013.00074

Ryals, A. J。, 王, J. X。, Polnaszek, K. L。, & Voss, J. L. (2015).
Hippocampal contribution to implicit conﬁguration memory ex-
pressed via eye movements during scene exploration. Hippo-
campus, 25(9), 1028–1041. https://doi.org/10.1002/hipo.22425
Ryan, J. D ., & 科恩, 氮. J. (2004). The nature of change detec-
tion and online representations of scenes. 实验杂志

网络神经科学

232

我

D
哦
w
n
哦
A
d
e
d

F
r
哦
米
H

t
t

:
/
/

d
我
r
e
C
t
.

米

我
t
.

e
d
你
n
e
n
A
r
t
我
C
e
–
p
d

我

F
/

4
1
2
1
7
1
8
6
6
7
5
4
n
e
n
_
A
_
0
0
1
2
0
p
d

乙
y
G
你
e
s
t

哦
n
0
9
S
e
p
e
米
乙
e
r
2
0
2
3

Simulated dynamics of hippocampal responses in oculomotor regions

心理学: Human Perception and Performance, 30(5), 988–1015.
https://doi.org/15462635

Ryan, J. D ., Althoff, 右. R。, Whitlow, S。, & 科恩, 氮. J. (2000). Amne-
sia is a deﬁcit in relational memory. 心理科学, 11(6),
454–461. https://doi.org/10.1111/1467-9280.00288

Ryan, J. D ., & 科恩, 氮. J. (2003). Evaluating the neuropsycho-
logical dissociation evidence for multiple memory systems. 齿轮-
nitive, Affective and Behavioral Neuroscience, 3(3), 168–185.
https://doi.org/10.3758/CABN.3.3.168

Ryan, J. D ., Leung, G。, Turk-Browne, 氮. B., & Hasher, L. (2007).
Assessment of age-related changes in inhibition and binding us-
ing eye movement monitoring. Psychology and Aging, 22(2),
239–250. https://doi.org/10.1037/0882-7974.22.2.239

Sato, N。, & Nakamura, K. (2006). Visual response properties of neu-
rons in the parahippocampal cortex of monkeys. https://doi.org/
10.1152/杰恩.01089.2002

Schmolesky, 中号. T。, 王, Y。, Hanes, D. P。, 汤普森, K. G。,
Leutgeb, S。, 沙尔, J. D ., & Leventhal, A. G. (2017). Signal timing
across the macaque visual system. 神经生理学杂志.
https://doi.org/10.1152/jn.1998.79.6.3272

Seelke, A. 中号. H。, Padberg,

J。, Disbrow, E., Purnell, S. M。,
Recanzone, G。, & Krubitzer, L. (2012). Topographic maps within
布罗德曼区 5 of macaque monkeys. 大脑皮层.
https://doi.org/10.1093/cercor/bhr257

沉, K., Bezgin, G。, Schirner, M。, 里特尔, P。, Everling, S。, &
McIntosh, A. 右. (2019). A macaque connectome for large-scale
network stimulations in TheVirtualBrain. Scientiﬁc Data, 6(1),
123. https://doi.org/10.1038/s41597-019-0129-z

沉, K., Bezgin, G。, Selvam, R。, McIntosh, A. R。, & Ryan, J. D. (2016).
An anatomical interface between memory and oculomotor sys-
特姆斯. 认知神经科学杂志, 28(11), 1772–1783.
https://doi.org/10.1162/jocn_a_01007

沉, K., Goulas, A。, Grayson, D. S。, Eusebio, J。, Gati, J. S。, Menon,
右. S。, . . . Everling, S. (2019). Exploring the limits of network topol-
ogy estimation using diffusion-based tractography and tracer
studies in the macaque cortex. 神经影像, 191, 81–92. https://
doi.org/10.1016/j.neuroimage.2019.02.018

史密斯, C. N。, 霍普金斯, 右. 奥。, & Squire, L. 右. (2006). 经验-
dependent eye movements, 意识, and hippocampus-
dependent memory. The Journal of Neuroscience: The Ofﬁcial
Journal of the Society for Neuroscience, 26(44), 11304–11312.
https://doi.org/10.1523/JNEUROSCI.3071-06.2006

史密斯, C. N。, & Squire, L. 右. (2017). When eye movements express
memory for old and new scenes in the absence of awareness and
independent of hippocampus. Learning and Memory. https://土井.
org/10.1101/lm.043851.116

Sobotka, S。, Nowicka, A。, & Ringo, J. L. (1997). Activity linked to ex-
ternally cued saccades in single units recorded from hippocam-
朋友, parahippocampal, and inferotemporal areas of macaques.
J Neurophysiol, 78(4), 2156–2163.

Spiegler, A。, 汉森, 乙. C. A。, Bernard, C。, McIntosh, A. R。, & Jirsa,
V. K. (2016). Selective activation of resting-state networks follow-
ing focal stimulation in a connectome-based network model of
the human brain. Eneuro, 3(5).

Stephan, K. E., Kamper, L。, Bozkurt, A。, Burns, G。, Young, M。, &
Kötter, 右. (2001). Advanced database methodology for the Col-
lation of Connectivity data on the Macaque brain (CoCoMac).
Philosophical Transactions of the Royal Society of London. Series
乙, Biological Sciences, 356, 1159–1186. https://doi.org/10.1098/
rstb.2001.0908

唐, Q., Burgalossi, A。, Ebbesen, C. L。, Sanguinetti-Scheck, J. 我。,
施密特, H。, Tukker, J. J。, . . . Brecht, 中号. (2016). Functional archi-
tecture of the rat parasubiculum. 神经科学杂志. https://
doi.org/10.1523/JNEUROSCI.3749-15.2016

Vincent, J. L。, Kahn, 我。, Van Essen, D. C。, & 巴克纳, 右. L. (2010).
Functional connectivity of the macaque posterior parahippocam-
pal cortex. 神经生理学杂志. https://doi.org/10.1152/
杰恩.00546.2009

Vlachos, 我。, Aertsen, A。, & Kumar, A. (2012). Beyond statistical sig-
niﬁcance: Implications of network structure on neuronal activ-
性. 公共科学图书馆计算生物学, 8(1). https://doi.org/10.1371/
journal.pcbi.1002311

Warren, D. E., Duff, 中号. C。, Tranel, D ., & 科恩, 氮. J. (2010).
Medial temporal lobe damage impairs representation of simple
刺激. Frontiers in Human Neuroscience, 4, 35. https://doi.org/
10.3389/fnhum.2010.00035

Wynn, J. S。, Ryan, J. D ., & Moscovitch, 中号. (2019). Effects of prior
knowledge on active vision and memory in younger and older
adults. 实验心理学杂志: General. https://土井.
org/10.1037/xge0000657

Yassa, 中号. A。, & Stark, C. 乙. L. (2011). Pattern separation in the
hippocampus. Trends in Neurosciences. https://doi.org/10.1016/
j.tins.2011.06.006

Yeung, L.-K., Olsen, 右. K., Bild-Enkin, H. 乙. P。, D’Angelo, 中号. C。,
Kacollja, A。, McQuiggan, D. A。, . . . Barense, 中号. D. (2017). 一个-
terolateral entorhinal cortex volume predicted by altered intra-
item conﬁgural processing. The Journal of Neuroscience. https://
doi.org/10.1523/JNEUROSCI.3664-16.2017

Yeung, L.-K., Olsen, 右. K., 洪, B., Mihajlovic, 五、, D’Angelo,
中号. C。, Kacollja, A。, . . . Barense, 中号. D. (2019). Object-in-
place memory predicted by anterolateral entorhinal cortex and
parahippocampal cortex volume in older adults. Journal of Cog-
nitive Neuroscience, 31(5), 711–729. https://doi.org/10.1162/
jocn_a_01385