The effects of dorsal Dentate Gyrus DREADD silencing on pattern
separation in rats using a location discrimination task.
Author: Jesse Groot Student number: 11012579
Datum: 17-06-2020 University: University of Amsterdam
Abstract
The hippocampus uses pattern separation during the memory consolidation process to minimalize the interference between different memories. Pattern separation of information in the hippocampus is highly associated with the Dentate Gyrus. Various studies have shown that DG lesions in rats result in a decreased ability to separate objects from each other with a similar but not dissimilar spatial location, indicating a decrease in pattern separation. In this research rats preformed a location discrimination task measuring the performance and behavioural changes by silencing DG neural activity. However instead of a DG lesion a DREADDs virus is used to accurately target the dorsal DG cells. The DREADDs virus can be activated during the testing phase by injecting clozapine N-oxide resulting in a inhabitation of the dorsal DG cells. Since no lesion was made neural activity was not be affected until the test period so no structural changes in the neural network could occur. Locomotion activity of the rats was quantified using a deep learning algorithm: DeepLabCut. This research showed no significant differences in location discrimination performance nor changes in behaviour, indicating no changes in pattern separation. However due to the low sample size this research is prone to unknown errors and given the low p-values (0.05 < p < 0.07) a replication of this study implementing changes for the shortcomings may change these findings.
Key words: Dentate Gyrus, Pattern Separation, DREADDs virus, LD task.
Introduction
The hippocampus is thought to be the centre pieces for encoding episodic, associative and contextual source memory (Witter, 2015; Yassa et al., 2011; Lemaire et al., 2012; Morris et al., 2012; Kenton et al., 2018). The hippocampus takes sensory information form events and stores it short term, when the input is frequently repeated storage is allocated to other cortices across the brain using ripples, causing long term memory (Staresina et al., 2015). To successfully storage memory the hippocampus depends on two complementary processes: pattern separation to minimize interference between memories and pattern completion to retrieve memories with partial input cues.
The two hippocampal regions most associated with pattern separation and pattern completion are the Dentate Gyrus (DG) and CA3. This is because the DG and CA3 show a stepwise transfer function that is able to detect small changes in the sensory input where in the CA1 increases in change of input shows similar a responses in graded fashion (Lacy et al., 2011). This makes the DG and CA3 able to perform nonlinear transformations (Guzowski, Knierim and Moser, 2004) which are needed for pattern separation and pattern completion (Yassa and Stark, 2011).
The function of the DG and CA3 in these processes are different. The DG is thought to play a major role in pattern separation and the CA3 in pattern completion (Neunuebel and Knierim, 2014).
According to this it is shown that the DG plays a crucial role in distinguishing stimuli which are located close to each other (Morris et al., 2012;Clelland et al., 2009;Gilbert, Kesner and Lee, 2001;Hunsaker and Kesner, 2008). In Morris et al. (2012) this was shown by using rats which had to perform a radial arm maze. The maze consisted of eight arms, of which one was fixed with a reward. The rats had to choose between the fixed arm and a second arm with a small or large separation between each other. Morris et al., (2012) showed that rats with a DG lesion made more mistakes during the small separation trials than in the control condition, indicating impairments in pattern separation between similar but not dissimilar locations.
Although DG lesions impair the ability to separate locations close to each other, this impairment does not bring the performance down to chance level (50%) indicating that the ability to separate locations still happens on a basal level. Lesions to the CA3 however, reduce the ability to separate locations to chance level (50%) this was proven using a spatial location-cued object recall task (Kesner, Hunsaker and Warthen, 2008). The CA3 contains place cells which fire for a specific location in space (i.e. place field) (O’Keefe, 1976). Because place cells code for different places in the environment an individual can track its location in its environment and store information to specific locations (Wilson and McNaughton, 1993). This explains why modulations within the CA3 result in differences in performance within spatial associative memory tasks (Leutgeb et al., 2007a; Nakashiba et al., 2008; Morris et al., 2012; Solstad, Yousif and Sejnowski, 2014;Le Duigou et al., 2014). The DG also contains place cells however these fire for up to four different locations in its environment (Leutgeb et al., 2007a) making it hard to store specific objects to specific locations. With the CA3 being the major output region of the DG (Rolls, 2013) and CA3 but not DG lesions impair the performance in a location discrimination task to chance level it is presumed that the DG is a key player for pattern separation of neural information to support CA3 storage.
The DG is the only region in the hippocampus where neurogenesis takes place. These new born granule cells have a high plasticity which is crucial for pattern separation (Trinchero et al., 2017; Nakashiba et al., 2012). As granule cells become older they lose plasticity and shift their function to pattern completion (Nakashiba et al., 2012). Several studies have shown that the DG play a role in differentiating multiple contexts from each other (McHugh et al., 2007;Tronel et al., 2012) which could be a key factor in its function of pattern separation. Old granule cells due to their low plasticity could code for the already known contexts in your life while new born cells change flow of information needed when new contexts occur. This also explains why the ability to store now memories decays with age since the neurogenesis decreases which hampers the function of pattern separation (Lichtenwalner et al., 2001;Yassa et al., 2011). The DG and CA3 both receive its major input through the perforant pathway deriving from the lateral entorhinal cortex (LEC) (Rolls, 2013;Kohara et al., 2014). The LEC shows spatial selectivity in presence of external stimuli (Deshmukh and Knierim, 2011) such as pictures or objects (Zhuc, Brown and Aggleton, 1995;Yeung et al., 2019). This information could be used by the DG to create an internal context of the outside world while CA3 place cells can provide spatial dimension to those objects.
In Lu et al. (2015) rats were used to measure the activity of CA3 place cells in 11 different rooms. Around 30% of the place cells did not activate in any of the rooms while approximately 60% of the place cells only activated in one to three of the 11 rooms (Lu et al., 2015) indicating place cell selectivity for specific contexts. The precise function of the DG on this phenomena however remains unclear. Although the CA3 pyramidal cell layer only perceives about 1-2 percent of the pyramidal cell layer input from the DG (excluding CA3 recurrent collaterals) these connections have a much stronger connectivity than the input through the perforant path (Rolls, 2013). The DG input has a large influence on CA3 spiking activity, lowering it significantly (McHugh et al., 2007) while increasing the amount of CA3 place cells firing (Sasaki et al., 2018).
It is clear that rats with a DG lesion show a decreased performance in spatial tasks with a high degree of spatial similarity (Gilbert, Kesner and Lee, 2001;Hunsaker and Kesner, 2008;Clelland et al., 2009;Morris et al., 2012). However according to what the CA3 activity is changed to cause this deterioration is not. When two objects are close to each other, one place cell could activate for both objects as they are in the same place field (see figure 1). In this case the DG can activate additional place cells to
code for each independent object resulting in less similarity between the firing responses and therefor makes it able to more correctly separate both items.
In this research further elaboration is done on the decreased performance in a spatial tasks with a high degree of spatial similarity as was shown in Gilbert et al., (2001), Hunsaker et al., (2008) and Morris et al., (2012). However instead
of lesioning the DG a DREADDs virus will be used to silence the DG cells to ensure that only the DG output is affected. A lesion can also damage surrounding areas
and gives the brain a window between the operation and testing period where structural changes may happen to adapt for the missing structure. This study would elaborate further on the function of the DG in pattern separation and give a clear method to selectively inhibit the DG cells which could be used in further studies.
To test the effect of DG cells on the pattern separation ability a hM4Di DREADDs virus was injected in the dorsal DG targeting excitatory cells. The rats performed a small and large location discrimination (LD) task in a singular environment (context) to study the LD ability and behaviour on nullifying DG output. During the task changes in performance and behaviour will be measured between the different conditions targeting trials to reach criterion, distance travelled and speed.
A adapted version of the LD task presented in Oomen et al., (2013) was used. However the rat would not perform the task from a fixed location but was able to freely move in a box. Much like presented in figure 1, two light squares would appear on the wall with a small (A – B) or a large (C – D) distance separation. the rat has to move to the left located stimulus and after 9 out of 10 correct responses the target stimulus would reverse to the other side restarting the learning process. Prior to the test conditions rats have received a clozapine N-oxide (CNO) injection to activate the DREADDs which inhibit the DG cells. As control the rats have received a Saline (VEH) injection instead which has no influence on the neural processes.
Since the DG signal should be important in minimalizing overlay in CA3 place cell signals between stimuli close to each other, behavioural changes in the small separation variant of the LD task should occur. In the large separation variant no difference should occur since place fields in the CA3 should do not overlay with each and therefor neural signals should not intervene with .
Negative changes in performance are expected during the small separation LD task when CNO is received compared to the VEH condition. This because the CNO injection will inhibit the DG output and this will cause higher levels of neural interference between the similar spatial locations. This will cause the rat to need more trials to reach the criteria of 9/10 correct responses. This also will increase the response latency of the rat since the task will increase in difficulty. However changes in locomotion of the rat should not occur since the path to either stimulus does not change and therefor the travelled distance should be equal. The large separation LD variant should not give any changes between the CNO and VEH condition since dissimilar locations do not show an increase neural interference and therefor difficulty will not increase.
Figure 1, The circles represent the place fields of two place cells in the CA3. The boxes (A-D) represent the different items in the environment. Object A and B would activate both place cells due to their similar spatial location object C and D do not.
Methods
Subjects
8 lister hooded rats (all male) were used, each 7 weeks old at arrival. Prior to the LD testing, the rats were trained five times per week for a total of 20 weeks. Where at the age of 27 weeks the LD testing took place. The rats were kept on a reversed day-light cycle and were food restricted while
maintaining weight above 85% of the normal ad libitum weight curve, no water restriction occurred.
Viral injection surgery
The rats went through viral injection surgery at the age of 18 weeks. The rats will be receive 2 subcutaneous injections of Buprenorphine (0.01-0.05 mg/kg s.c, analgesic) and Meloxicam (2mg/kg, analgesic) 20 – 30 minutes prior to anaesthesia. Thereafter, the rat is anaesthetized by inhaling a mixture of 100% oxygen with isoflurane (Induction dose 3.0% - 5.0%, Maintenance dose 1.0% - 3.0%). After incision Lidocaine base (10-20mg lidocaine base) is applied as a local anaesthetic on the periost. The DREADD virus pAAV-CaMKIIa-hM4D(Gi)-mCherry (0.41 microliter, source: viral vector facility, Neuroscience Center Zurich) was injected bilateral using a nanoject (Drummond Scientific Company, Broomall, PA) at the injection sites 2x DG op bregma -2.7 AP, +-1,2 ML, -3.7 DV and 2x DG op bregma -3.7 AP, +-2,0 ML, -3.1 DV. The DREADD virus contains a tag (mCherry) which will light up against green light. This will be used to check virus expression past death.
Apparatus
For the LD task a box of 52x50x40 cm was used with a monitor at the front wall of 52x40 cm. Before the monitor are 8 wells with a width of 6 cm leaving 2 cm to the outer walls and a ninth well in the middle of the box. Each well has a nose poke which released a microliter 15% sucrose water if the correct response was given at the end of a trial. If an incorrect response was given, the rat was punished using light on the back wall that shone brightly for 3 seconds. If the correct response was given the next trial could be started after an interval of 10 seconds. When the incorrect stimulus was chosen the next trial could start 12 seconds after the punishing occurred, resulting in 15 seconds between the incorrect response and the start of the new trial. Depending on the condition (large or small LD task) that was tested the monitor will be used to display 2 stimuli above well 2 and 7 (30 cm separation) or 4 and 5 (6 cm separation). The behaviour of the rats was filmed with a camera placed directly above the box for further analysis.
Procedure LD task
The four conditions: large LD CNO, large LD VEH, small LD CNO and small LD VEH were done in four days (one condition/day). The rats received an injection of clozapine N-oxide/CNO (activating DREADDs virus) or Saline/VEH (control) 30 minutes prior to the LD task. Thereafter the rats were placed in the box in a dark room where they performed a condition for a length of 60 minutes. A trial could be initiated when a light shone upon the central well. Upon pressing the nose poke the two stimuli appeared above the wells. Where the rat has to response by pressing the nose poke located under the correct stimulus. The correct response stimulus would be either at the left or right side and change when 9 out of 10 trials are correct. The first three reversals were be used for statistical testing.
Video tracking with DeepLabCut
Figure 2, frame from DeepLabCut where every tracked body part is displayed using the algorithm. The lines between the dots connect the datapoints that will be used together.
To analyse the behaviour of the rats a program called DeepLabCut was used. DeepLabCut is a deep learning algorithm that takes labelled frames as input and learns an algorithm to perform similar labelling to the remaining frames. For every rat 60 frames per condition were labelled (240 in total), focussing on seven points: Central well, Right corner of the box, left and right ear and the upper, middle and lower body of the rat. The DeepLabCut output contained a table where each frame has his own array. This array contained the X and Y coordinate of every point and a likelihood between 0-1 indicating the legibility of the coordinate.
The camera was located in different positions over the various sessions, this caused the box to appear in different sizes on the videos. Because of between two different sessions the distance between 2 fixed coordinates would differ in actual distance. This because the two coordinates within the frame would have the same distance from each other but since the box is smaller in one session compared the other the distance presented by those coordinates is larger. To correctly calculate the distance between coordinates a sum of all right corner and centre well coordinates was used if a likelihood > 0.9 was given to the coordinate. Since the box was 52 by 50 cm the difference between these coordinates would be sqr(26^2 + 25^2) = 36 cm. this was used to calculate the relative distance within each trial.
To track the rat during each trial, the start and finish time was used to extract the corresponding frames from DeepLabCut. Within every frame the coordinates with a likelihood >0.7 were used to locate the rats location. The middle body was taken as reverence point of the rats body. If the middle body had a likelihood < 0.7 and the upper and lower body had a likelihood of > 0.7 the mean between those two coordinates was taken instead. If the frame did not meet these criteria the frame is skipped to the next one.
To accurately calculate the distance travelled within each trial, the sum of the differences between each consecutive coordinate was taken (sqr((abs|X1-X2|)^2+( abs|Y1-Y2|)^2)). However when the rat sat still, the small errors within the coordinates would slightly change the location of the rat over the frames creating the illusion that distance was travelled. Therefor a new coordinate was used when the X or Y coordinate changed at least 2.4 cm from his previous position. This would be calculated by the distance between start point and right corner divided by 15.
Data Analysis
The LD box registered the trials to reach a reversal. The response latency within each trial was obtained by calculating the difference in time between when the central nose poke was triggered and a nose poke beneath the given stimuli. At last the distance travelled was obtained by usage of the DeepLabCut program.
A Factorial Repeated Measures ANOVA (Two-way repeated measures ANOVA) was used on each of the different data points to calculate the interaction effect of the DREADDs virus effects and the location distance between the stimuli on trials to reach criterion, travelled distance and response latency. As post-hoc a pairwise comparison was done if needed. For all tests a p-value of < 0.05 is used.
Results
The bilateral virus expression to the dorsal DG where checked extracting brain slices between 2.7–3.7 mm posterior to bregma. Six out of eight rats showed normal levels of expression but rat 3 showed no virus expression and rat 5 only in the right hemisphere and are therefore excluded from the analysis (see appendix).
LD Ability
Differences in LD ability between the CNO and VEH condition were determined using a factorial repeated measures ANOVA. This was based on the mean trials to reach criterion within the small and large LD conditions. As can be seen in figure 3A, no significant differences were found between the conditions (numDF= 1, denDF=15, F-value = 4.04630, P-value = 0.0626).
Behavioural analysis
The rats behaviour was analysed based on their travelled distance and response latency within each trial. Due to a mistake in preparation, the first session was filmed with an insignificant amount of light. This caused the recordings to be too dark to accurately locate the rats coordinates using DeepLabCut which resulted in low likelihood numbers for these coordinates and thus location of the rats (see figure
4B). Middle body coordinates were roughly collected, but suffered from more likelihood variation compared to later sessions (figure 4A). The data was slightly influenced by these errors. Due to an error in likelihood extraction no coordinates could be obtained in rat 4 during the large LD CNO condition. This forced exclusion of rat 4 within the distance variant of behavioural analysis
Differences in response latency and travelled distance between the CNO and VEH condition were determined using a factorial repeated measures ANOVA. This was based on the mean of every variable within each condition. As can be seen in figure 3B and 3C no significant differences were found in either response latency or distance (response latency: numDF= 1, denDF=15, F-value = 4.34074, P-value = 0.0547; distance: numDF= 1, denDF=12, F-value = 1.1175, P-value = 0.3113).
Discussion
Figure 3, mean (±SE) of the various tested variables within the large (L) and small (S) separation LD task with the DG cells being silenced by CNO or unaffected by VEH. (A) Mean trials needed to correctly respond to the given stimuli. (B) Mean response latency in seconds needed for reaction to the stimuli. (C) Mean distance traveled in centimeters during the trials
Figure 4, Likelihood numbers per frame generated by DeepLabCut. (A) Likelihood numbers within a video with sufficient lighting. (B) Likelihood numbers within a video with insufficient lightning.
In this study rats were used to investigate the influence of the dorsal DG on pattern separation using a large and small separation LD task. Prior to the task the rats received a CNO injection to inhibit the dorsal DG or a VEH injection as control. The results showed that inhibition of the dorsal DG did not significantly lower the LD ability in either the large or small separation variant. Likewise, no significant results were found within the behavioural expression of the rat. Between the conditions neither distance travelled nor response latency significantly changed within the trials. It was there for concluded that dorsal DG inhibition does not affect the pattern separation nor the behaviour of the rats within a LD task.
This however does not correspond with similar research in which the dorsal DG was completely lesioned. In Morris et al., (2012) rats preformed a spatial reference memory task including place learning for large and small separated items a study comparable to this research. They however, found a significant difference in the rats ability to separate small separated items from each other. The difference between our researches can be explained by findings from Clelland et al., (2009) where they only blocked the neurogenesis within the DG. This also resulted in a decrease of the LD ability to separate similar but not dissimilar lactations thus influencing pattern separation. The current research only targeted the dorsal DG cells present at time of the DREADDs virus injection. Since the dorsal DG was not affected between surgery and the testing phase, neurogenesis had been unaffected and the new born cells in the dorsal DG were not inhibited by the CNO injection.
In McDonald and Wojtowicz et al., (2005) was shown that rats between the age of 5.4 to 14 weeks generate around 5000 to 10000 new born cells a day, while at the age of 12 months this is reduced to 1000 (McDonald and Wojtowicz, 2005). Since there is an interval of 9 weeks between the DREADDs virus injection and the testing period (18 – 27 weeks old) it can be concluded that there are approximately 63 to 315/630 thousand new granule cells during the testing phase unaffected by the DREADDs virus. As was previously mentioned, many presume that new born granule cells play a role in pattern separation whilst old granule cells play a role in pattern completion (Nakashiba et al., 2012). his could explain why no significant change in the LD ability was found.
For the behavioral analysis the response latency was expected to change when the LD ability was affected. However since those are not significant, the non-significant findings in response latency are anticipated. The distance travelled within each trial was expected to stay the same in either situation and this was the case. Since behavior of the rat stayed the same during the conditions it can be presumed that no external functions beside the DG inhibition could have affected the LD ability.
However it has to be pointed out that this research used a small sample size which makes the results prone for undetectable errors. For example, the LD ability was measured within each condition by taking the trials to reach criterion of the first three reversals. Since only three reversals were used, one unintentional mistake from the rat could drastically increase the mean within one condition. Due to the low sample size intrinsic motivation of the rat to do the task might also influence the data points. Together with the results of LD ability and latency response being between 0.05 < p <0.07 a larger sample size in this study might still show significant results.
In further research this experiment should be repeated while simultaneously deactivating the neurogenesis as is done in Clelland et al., (2009). This will reveal whether new DG cells are responsible for the pattern separation. If results cohere with earlier lesion studies, contextual changes within the LD task should be added. By adding a contextual element, the function of the DG in pattern separation between various contexts could cause difficulties between largely separated conditions as well. As was mentioned in Leutgeb et al., (2007b), CA3 place cells have a higher spatial correlation between contexts than DG cells. Objects in different contexts on the same spatial location could therefor intervene with each other in the same way displayed in this research in the small separation variant. This could therefor increase the knowledge of the DG in the pattern separation mechanism.
In short, this research did not find any changes in LD ability nor behaviour by silencing dorsal DG cells and therefor pattern separation is unaffected. However, this could be caused by neurogenesis which generated a large amount of new born cells between the DREADDs virus
injection and testing phase. Another explanation could be the low sample size which made this research prone for undetectable errors.
Literature List:
Clelland, C. D. et al. (2009) ‘A functional role for adult hippocampal neurogenesis in spatial pattern separation’, Science, 325(5937), pp. 210–213. doi: 10.1126/science.1173215.
Deshmukh, S. S. and Knierim, J. J. (2011) ‘Representation of non-spatial and spatial information in the lateral entorhinal cortex’, Frontiers in Behavioral Neuroscience, (OCTOBER). doi:
10.3389/fnbeh.2011.00069.
Le Duigou, C. et al. (2014) ‘Recurrent synapses and circuits in the CA3 region of the hippocampus: An associative network’, Frontiers in Cellular Neuroscience. Frontiers Media S.A., 7(JAN). doi:
10.3389/fncel.2013.00262.
Gilbert, P. E., Kesner, R. P. and Lee, I. (2001) ‘Dissociating hippocampal subregions: A double
dissociation between dentate gyrus and CA1’, Hippocampus. John Wiley & Sons, Ltd, 11(6), pp. 626– 636. doi: 10.1002/hipo.1077.
Guzowski, J. F., Knierim, J. J. and Moser, E. I. (2004) ‘Ensemble dynamics of hippocampal regions CA3 and CA1’, Neuron. Cell Press, pp. 581–584. doi: 10.1016/j.neuron.2004.11.003.
Hunsaker, M. R. and Kesner, R. P. (2008) ‘Evaluating the differential roles of the dorsal dentate gyrus, dorsal CA3, and dorsal CA1 during a temporal ordering for spatial locations task’, Hippocampus. John Wiley & Sons, Ltd, 18(9), pp. 955–964. doi: 10.1002/hipo.20455.
Kenton, J. A. et al. (2018) ‘Cortico-hippocampal GluN2B is essential for efficient visual-spatial
discrimination learning in a touchscreen paradigm’, Neurobiology of Learning and Memory. Academic Press Inc., 156, pp. 60–67. doi: 10.1016/j.nlm.2018.10.011.
Kesner, R. P., Hunsaker, M. R. and Warthen, M. W. (2008) ‘The CA3 Subregion of the Hippocampus Is Critical for Episodic Memory Processing by Means of Relational Encoding in Rats’, Behavioral Neuroscience, 122(6), pp. 1217–1225. doi: 10.1037/a0013592.
Kohara, K. et al. (2014) ‘Cell type-specific genetic and optogenetic tools reveal hippocampal CA2 circuits’, Nature Neuroscience. Nature Publishing Group, 17(2), pp. 269–279. doi: 10.1038/nn.3614. Lacy, J. W. et al. (2011) ‘Distinct pattern separation related transfer functions in human CA3/dentate and CA1 revealed using highresolution fMRI and variable mnemonic similarity’, Learning and Memory. Cold Spring Harbor Laboratory Press, 18(1), pp. 15–18. doi: 10.1101/lm.1971111. Lemaire, V. et al. (2012) ‘Long-lasting plasticity of hippocampal adult-born neurons’, Journal of Neuroscience. Society for Neuroscience, 32(9), pp. 3101–3108. doi: 10.1523/JNEUROSCI.4731-11.2012.
Leutgeb, J. K. et al. (2007a) ‘Pattern separation in the dentate gyrus and CA3 of the hippocampus’, Science, 315(5814), pp. 961–966. doi: 10.1126/science.1135801.
Leutgeb, J. K. et al. (2007b) ‘Pattern separation in the dentate gyrus and CA3 of the hippocampus’, Science. American Association for the Advancement of Science, 315(5814), pp. 961–966. doi: 10.1126/science.1135801.
Lichtenwalner, R. J. et al. (2001) ‘Intracerebroventricular infusion of insulin-like growth factor-I ameliorates the age-related decline in hippocampal neurogenesis’, Neuroscience. Pergamon, 107(4), pp. 603–613. doi: 10.1016/S0306-4522(01)00378-5.
Lu, L. et al. (2015) ‘Topography of Place Maps along the CA3-to-CA2 Axis of the Hippocampus’, Neuron. Cell Press, 87(5), pp. 1078–1092. doi: 10.1016/j.neuron.2015.07.007.
McDonald, H. Y. and Wojtowicz, J. M. (2005) ‘Dynamics of neurogenesis in the dentate gyrus of adult rats’, Neuroscience Letters. Elsevier, 385(1), pp. 70–75. doi: 10.1016/j.neulet.2005.05.022.
McHugh, T. J. et al. (2007) ‘Dentate gyrus NMDA receptors mediate rapid pattern separation in the hippocampal network’, Science. American Association for the Advancement of Science, 317(5834), pp. 94–99. doi: 10.1126/science.1140263.
Morris, A. M. et al. (2012) ‘Selective lesions of the dentate gyrus produce disruptions in place learning for adjacent spatial locations’, Neurobiology of Learning and Memory, 97(3), pp. 326–331. doi: 10.1016/j.nlm.2012.02.005.
Nakashiba, T. et al. (2008) ‘Transgenic inhibition of synaptic transmission reveals role of CA3 output in hippocampal learning’, Science, 319(5867), pp. 1260–1264. doi: 10.1126/science.1151120.
Nakashiba, T. et al. (2012) ‘Young dentate granule cells mediate pattern separation, whereas old granule cells facilitate pattern completion’, Cell, 149(1), pp. 188–201. doi: 10.1016/j.cell.2012.01.046. Neunuebel, J. P. and Knierim, J. J. (2014) ‘CA3 retrieves coherent representations from degraded input: Direct evidence for CA3 pattern completion and dentate gyrus pattern separation’, Neuron. Cell Press, 81(2), pp. 416–427. doi: 10.1016/j.neuron.2013.11.017.
O’Keefe, J. (1976) ‘Place units in the hippocampus of the freely moving rat’, Experimental Neurology. Academic Press, 51(1), pp. 78–109. doi: 10.1016/0014-4886(76)90055-8.
Rolls, E. T. (2013) ‘The mechanisms for pattern completion and pattern separation in the
hippocampus’, Frontiers in Systems Neuroscience. Frontiers, p. 74. doi: 10.3389/fnsys.2013.00074. Sasaki, T. et al. (2018) ‘Dentate network activity is necessary for spatial working memory by supporting CA3 sharp-wave ripple generation and prospective firing of CA3 neurons’, Nature Neuroscience. Nature Publishing Group, 21(2), pp. 258–269. doi: 10.1038/s41593-017-0061-5. Solstad, T., Yousif, H. N. and Sejnowski, T. J. (2014) ‘Place Cell Rate Remapping by CA3 Recurrent Collaterals’, PLoS Computational Biology. Public Library of Science, 10(6). doi:
10.1371/journal.pcbi.1003648.
Staresina, B. P. et al. (2015) ‘Hierarchical nesting of slow oscillations, spindles and ripples in the human hippocampus during sleep’, Nature Neuroscience. Nature Publishing Group, 18(11), pp. 1679– 1686. doi: 10.1038/nn.4119.
Trinchero, M. F. et al. (2017) ‘High Plasticity of New Granule Cells in the Aging Hippocampus’, Cell Reports. Elsevier B.V., 21(5), pp. 1129–1139. doi: 10.1016/j.celrep.2017.09.064.
Tronel, S. et al. (2012) ‘Adult-born neurons are necessary for extended contextual discrimination’, Hippocampus, 22(2), pp. 292–298. doi: 10.1002/hipo.20895.
Wilson, M. A. and McNaughton, B. L. (1993) ‘Dynamics of the hippocampal ensemble code for space’, Science, 261(5124), pp. 1055–1058. doi: 10.1126/science.8351520.
Witter, M. P. (no date) ‘Hippocampus and Related Structures What Is the Hippocampus and What Are Related Structures?’ doi: 10.1016/B978-0-08-097086-8.51008-3.
Yassa, M. A. et al. (2011) ‘Pattern separation deficits associated with increased hippocampal CA3 and dentate gyrus activity in nondemented older adults’, Hippocampus, 21(9), pp. 968–979. doi:
Yassa, M. A. and Stark, C. E. L. (2011) ‘Pattern separation in the hippocampus’, Trends in Neurosciences. Elsevier Current Trends, pp. 515–525. doi: 10.1016/j.tins.2011.06.006.
Yeung, L. K. et al. (2019) ‘Object-in-place memory predicted by anterolateral entorhinal cortex and parahippocampal cortex volume in older adults’, Journal of Cognitive Neuroscience. MIT Press Journals, 31(5), pp. 711–729. doi: 10.1162/jocn_a_01385.
Zhu, X. O., Brown, M. W. and Aggleton, J. P. (1995) ‘Neuronal Sianallina of Information Imoortant to Visual Recognition Memory in Rat Rhinal aid Neighbouring Cortices’, ‐ European Journal of
Neuroscience, 7(4), pp. 753–765. doi: 10.1111/j.1460-9568.1995.tb00679.x.
Bijlage:
Figure 5, Brain slices between 2.7–3.7 mm posterior to bregma from rats 1-8 with in red the DREADDs virus mCherry tag. Rats 1, 2, 4, 6, 7 and 8 showed bilateral virus expression. Rat 3 showed no virus expression and rat 5 showed virus expression only in the right hemisphere.
