-------------------------------------------------------------------- revision 15.1 locked by: ajay; date: 2000/02/22 15:39:18; author: ajay; state: Exp; lines: +1 -1 simulation testing disambiguation. EC2pyr input used for disambiguating 2 forked sequences. Low learning rate. Does not work well when both sequences are learnt. Membrane parameters set so that tauM is large (Cm=3 uF/cm2, Rm=2ohm-m2). -------------------------------------------------------------------- revision 14.1 locked by: ajay; date: 2000/01/17 15:02:37; author: ajay; state: Exp; lines: +1 -1 batch run varying recall_exc_wt_mult from 1 - 10 (this factor is multiplied by pyr_ampa_wt of all synapses that underwent LTP, just prior to recall, step size 0.25), and EC_ILN_depol from 0 - 0.18 nA (step 0.03 nA) which injects current only into ILN neurons of pat2-5 of sequence 1 during recall. recall_gaba_wt_mult was 0.6 throughout. Data saved in bat_1_14.out -------------------------------------------------------------------- revision 13.1 locked by: ajay; date: 2000/01/13 14:32:33; author: ajay; state: Exp; lines: +1 -1 batch run varying EC_ILN_depol (current step into all bg neurons and all ILN cells in pat2-5 of sequence 1) from 0 to 0.18 nA (step 0.02 nA = about 1 mV depol step) and recall_gaba_wt_mult (factor which is multiplied by pyr_gabaa_wt and pyr_gabab_wt prior to recall phase) from 0.1 to 1 (step 0.1). recall_exc_wt_mult was set to 2. Data saved in bat_1_12.out. -------------------------------------------------------------------- revision 12.1 locked by: ajay; date: 2000/01/12 14:52:34; author: ajay; state: Exp; lines: +1 -1 batch run varying learn_rate and EC_ILN_depol (depol current to simulate EC input to ILN cells of sequence 1 and all bg neurons). Presented 2 sequences to network (pat1-5, pat1,6-9) and recalled seq. 1. Data saved in bat_1_11.out. -------------------------------------------------------------------- revision 11.1 locked by: ajay; date: 2000/01/10 18:41:12; author: ajay; state: Exp; lines: +1 -1 state of simulator used to generate figure of perfect recall in sparse network (p=0.1) for book chapter. -------------------------------------------------------------------- revision 10.1 locked by: ajay; date: 2000/01/03 16:49:54; author: ajay; state: Exp; lines: +6 -4 batch run to examine effect of varying p-p learnrate (70-140) and depol current into ILN cells (0.1 - 0.15 nA) that simulated biasing EC input that would serve to disambiguate overlapping sequences. gabaA and gabaB inhibition were reduced during the recall phase to 1% of the values used in the learning phase. -------------------------------------------------------------------- revision 9.1 locked by: ajay; date: 1999/12/24 13:59:48; author: ajay; state: Exp; lines: +1 -1 batch run with steady depol to bg neurons. Initial Vm varied from -65 to -50 mV. data saved in bat_12_21.out. pyr_gabaa_wt = 0.1. -------------------------------------------------------------------- revision 8.1 locked by: ajay; date: 1999/12/20 14:24:12; author: ajay; state: Exp; lines: +1 -2 USELESS!!!! Had pyr_gabaa_wt set to 1.0 instead of 0.1 in these simulations by mistake. batch run with staedy depol to bg neurons instead of triangular waveform. steady holding potentials varied from -65 to -50 mV. data saved in bat_12_15.out. -------------------------------------------------------------------- revision 7.1 locked by: ajay; date: 1999/12/14 19:16:26; author: ajay; state: Exp; lines: +1 -1 batch run looking at effect of varying pyr_gabaa_wt. Data in file bat_12_14.out (here only 0.04 used, but varied in previous batch runs whose data in bat_12_10.out and bat_12_11.out). -------------------------------------------------------------------- revision 6.1 locked by: ajay; date: 1999/12/10 14:34:23; author: ajay; state: Exp; lines: +1 -2 batch run varying bg_depol_current from 0.14 to 0.25 nA and bg_depol_ramp_period from 13 to 17 ms. 50 diff. values of ranseed for each set of params. data saved in bat_12_3.out -------------------------------------------------------------------- revision 5.1 locked by: ajay; date: 1999/11/16 22:00:20; author: ajay; state: Exp; lines: +2 -1 batch run to look at effects of varying plateau depolarization. FAILED!! Since I didn't call 'add_sawtooth_depol' in the initialization phase of 'lrs', the values of bg_depol_current and bg_depol_plateau_period were not changed form the values defined in constants.g (0.18nA and 10ms resp.). Therefore, all runs had same params except for ranseed. -------------------------------------------------------------------- revision 4.1 locked by: ajay; date: 1999/11/09 13:57:17; author: ajay; state: Exp; lines: +12 -10 batch run. Varied pyr_gabaa_wt from 0.06 to 0.14 (+0.02). For each pyr_gabaa_wt value did 200 runs with different ranseed values. -------------------------------------------------------------------- revision 3.1 locked by: ajay; date: 1999/11/03 18:47:55; author: ajay; state: Exp; lines: +16 -7 batch files looking at variability in bg firing across patterns in a sequence. Only ranseed was varied in the batch run. -------------------------------------------------------------------- revision 2.1 locked by: ajay; date: 1999/09/24 18:06:45; author: ajay; state: Exp; lines: +1 -1 perfect recall, p=0.1, GW's full cell, AHP=80, gabaA and B inhibition. NOTE: SAME AS REV 1.3. FORCED SAVE WITH REV 2.1 SO THAT ALL FILES HAVE SAME REV NUMBER! -------------------------------------------------------------------- revision 1.3 locked by: ajay; date: 1999/09/24 17:53:17; author: ajay; state: Exp; lines: +5 -2 perfect recall with p->p = 0.1. GW's model cell. GABAa and b inh. N=200. NOTE: FILES THAT WERE NOT CHANGED ARE STILL REV. 1.2 -------------------------------------------------------------------- revision 1.2 locked by: ajay; date: 1999/07/22 20:05:14; author: ajay; state: Exp; lines: +2 -1 perfect recall for sparse network, p=0.2, GW's cell with high AHP. No inhibition reqd. for perfect recall. ---------------------------------------------------------------------- revision 1.1 date: 1999/06/29 18:12:38; author: ajay; state: Exp; Initial revision unique inhibition scheme to turn off cells that have just fired ----------------------------------------------------------------------